Online semi-supervised learning applied to an automated insect pest monitoring system
The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the...
Ausführliche Beschreibung
Autor*in: |
Rustia, Dan Jeric Arcega [verfasserIn] |
---|
Format: |
E-Artikel |
---|---|
Sprache: |
Englisch |
Erschienen: |
2021transfer abstract |
---|
Schlagwörter: |
---|
Umfang: |
17 |
---|
Übergeordnetes Werk: |
Enthalten in: The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) - Raggi, M. ELSEVIER, 2015, San Diego, Calif |
---|---|
Übergeordnetes Werk: |
volume:208 ; year:2021 ; pages:28-44 ; extent:17 |
Links: |
---|
DOI / URN: |
10.1016/j.biosystemseng.2021.05.006 |
---|
Katalog-ID: |
ELV054593840 |
---|
LEADER | 01000caa a22002652 4500 | ||
---|---|---|---|
001 | ELV054593840 | ||
003 | DE-627 | ||
005 | 20230626040455.0 | ||
007 | cr uuu---uuuuu | ||
008 | 210910s2021 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.1016/j.biosystemseng.2021.05.006 |2 doi | |
028 | 5 | 2 | |a /cbs_pica/cbs_olc/import_discovery/elsevier/einzuspielen/GBV00000000001447.pica |
035 | |a (DE-627)ELV054593840 | ||
035 | |a (ELSEVIER)S1537-5110(21)00106-9 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
082 | 0 | 4 | |a 630 |q VZ |
082 | 0 | 4 | |a 640 |q VZ |
082 | 0 | 4 | |a 320 |q VZ |
082 | 0 | 4 | |a 630 |a 640 |a 610 |q VZ |
100 | 1 | |a Rustia, Dan Jeric Arcega |e verfasserin |4 aut | |
245 | 1 | 0 | |a Online semi-supervised learning applied to an automated insect pest monitoring system |
264 | 1 | |c 2021transfer abstract | |
300 | |a 17 | ||
336 | |a nicht spezifiziert |b zzz |2 rdacontent | ||
337 | |a nicht spezifiziert |b z |2 rdamedia | ||
338 | |a nicht spezifiziert |b zu |2 rdacarrier | ||
520 | |a The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. | ||
520 | |a The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. | ||
650 | 7 | |a insect monitoring |2 Elsevier | |
650 | 7 | |a feature extraction |2 Elsevier | |
650 | 7 | |a integrated pest management |2 Elsevier | |
650 | 7 | |a semi-supervised learning |2 Elsevier | |
650 | 7 | |a image recognition |2 Elsevier | |
700 | 1 | |a Lu, Chen-Yi |4 oth | |
700 | 1 | |a Chao, Jun-Jee |4 oth | |
700 | 1 | |a Wu, Ya-Fang |4 oth | |
700 | 1 | |a Chung, Jui-Yung |4 oth | |
700 | 1 | |a Hsu, Ju-Chun |4 oth | |
700 | 1 | |a Lin, Ta-Te |4 oth | |
773 | 0 | 8 | |i Enthalten in |n Academ. Press |a Raggi, M. ELSEVIER |t The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) |d 2015 |g San Diego, Calif |w (DE-627)ELV018374581 |
773 | 1 | 8 | |g volume:208 |g year:2021 |g pages:28-44 |g extent:17 |
856 | 4 | 0 | |u https://doi.org/10.1016/j.biosystemseng.2021.05.006 |3 Volltext |
912 | |a GBV_USEFLAG_U | ||
912 | |a GBV_ELV | ||
912 | |a SYSFLAG_U | ||
912 | |a SSG-OLC-PHA | ||
912 | |a GBV_ILN_20 | ||
912 | |a GBV_ILN_22 | ||
912 | |a GBV_ILN_24 | ||
912 | |a GBV_ILN_30 | ||
912 | |a GBV_ILN_40 | ||
912 | |a GBV_ILN_65 | ||
912 | |a GBV_ILN_70 | ||
951 | |a AR | ||
952 | |d 208 |j 2021 |h 28-44 |g 17 |
author_variant |
d j a r dja djar |
---|---|
matchkey_str |
rustiadanjericarcegaluchenyichaojunjeewu:2021----:nieeiuevsderigpletaatmtdnet |
hierarchy_sort_str |
2021transfer abstract |
publishDate |
2021 |
allfields |
10.1016/j.biosystemseng.2021.05.006 doi /cbs_pica/cbs_olc/import_discovery/elsevier/einzuspielen/GBV00000000001447.pica (DE-627)ELV054593840 (ELSEVIER)S1537-5110(21)00106-9 DE-627 ger DE-627 rakwb eng 630 VZ 640 VZ 320 VZ 630 640 610 VZ Rustia, Dan Jeric Arcega verfasserin aut Online semi-supervised learning applied to an automated insect pest monitoring system 2021transfer abstract 17 nicht spezifiziert zzz rdacontent nicht spezifiziert z rdamedia nicht spezifiziert zu rdacarrier The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition Elsevier Lu, Chen-Yi oth Chao, Jun-Jee oth Wu, Ya-Fang oth Chung, Jui-Yung oth Hsu, Ju-Chun oth Lin, Ta-Te oth Enthalten in Academ. Press Raggi, M. ELSEVIER The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) 2015 San Diego, Calif (DE-627)ELV018374581 volume:208 year:2021 pages:28-44 extent:17 https://doi.org/10.1016/j.biosystemseng.2021.05.006 Volltext GBV_USEFLAG_U GBV_ELV SYSFLAG_U SSG-OLC-PHA GBV_ILN_20 GBV_ILN_22 GBV_ILN_24 GBV_ILN_30 GBV_ILN_40 GBV_ILN_65 GBV_ILN_70 AR 208 2021 28-44 17 |
spelling |
10.1016/j.biosystemseng.2021.05.006 doi /cbs_pica/cbs_olc/import_discovery/elsevier/einzuspielen/GBV00000000001447.pica (DE-627)ELV054593840 (ELSEVIER)S1537-5110(21)00106-9 DE-627 ger DE-627 rakwb eng 630 VZ 640 VZ 320 VZ 630 640 610 VZ Rustia, Dan Jeric Arcega verfasserin aut Online semi-supervised learning applied to an automated insect pest monitoring system 2021transfer abstract 17 nicht spezifiziert zzz rdacontent nicht spezifiziert z rdamedia nicht spezifiziert zu rdacarrier The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition Elsevier Lu, Chen-Yi oth Chao, Jun-Jee oth Wu, Ya-Fang oth Chung, Jui-Yung oth Hsu, Ju-Chun oth Lin, Ta-Te oth Enthalten in Academ. Press Raggi, M. ELSEVIER The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) 2015 San Diego, Calif (DE-627)ELV018374581 volume:208 year:2021 pages:28-44 extent:17 https://doi.org/10.1016/j.biosystemseng.2021.05.006 Volltext GBV_USEFLAG_U GBV_ELV SYSFLAG_U SSG-OLC-PHA GBV_ILN_20 GBV_ILN_22 GBV_ILN_24 GBV_ILN_30 GBV_ILN_40 GBV_ILN_65 GBV_ILN_70 AR 208 2021 28-44 17 |
allfields_unstemmed |
10.1016/j.biosystemseng.2021.05.006 doi /cbs_pica/cbs_olc/import_discovery/elsevier/einzuspielen/GBV00000000001447.pica (DE-627)ELV054593840 (ELSEVIER)S1537-5110(21)00106-9 DE-627 ger DE-627 rakwb eng 630 VZ 640 VZ 320 VZ 630 640 610 VZ Rustia, Dan Jeric Arcega verfasserin aut Online semi-supervised learning applied to an automated insect pest monitoring system 2021transfer abstract 17 nicht spezifiziert zzz rdacontent nicht spezifiziert z rdamedia nicht spezifiziert zu rdacarrier The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition Elsevier Lu, Chen-Yi oth Chao, Jun-Jee oth Wu, Ya-Fang oth Chung, Jui-Yung oth Hsu, Ju-Chun oth Lin, Ta-Te oth Enthalten in Academ. Press Raggi, M. ELSEVIER The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) 2015 San Diego, Calif (DE-627)ELV018374581 volume:208 year:2021 pages:28-44 extent:17 https://doi.org/10.1016/j.biosystemseng.2021.05.006 Volltext GBV_USEFLAG_U GBV_ELV SYSFLAG_U SSG-OLC-PHA GBV_ILN_20 GBV_ILN_22 GBV_ILN_24 GBV_ILN_30 GBV_ILN_40 GBV_ILN_65 GBV_ILN_70 AR 208 2021 28-44 17 |
allfieldsGer |
10.1016/j.biosystemseng.2021.05.006 doi /cbs_pica/cbs_olc/import_discovery/elsevier/einzuspielen/GBV00000000001447.pica (DE-627)ELV054593840 (ELSEVIER)S1537-5110(21)00106-9 DE-627 ger DE-627 rakwb eng 630 VZ 640 VZ 320 VZ 630 640 610 VZ Rustia, Dan Jeric Arcega verfasserin aut Online semi-supervised learning applied to an automated insect pest monitoring system 2021transfer abstract 17 nicht spezifiziert zzz rdacontent nicht spezifiziert z rdamedia nicht spezifiziert zu rdacarrier The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition Elsevier Lu, Chen-Yi oth Chao, Jun-Jee oth Wu, Ya-Fang oth Chung, Jui-Yung oth Hsu, Ju-Chun oth Lin, Ta-Te oth Enthalten in Academ. Press Raggi, M. ELSEVIER The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) 2015 San Diego, Calif (DE-627)ELV018374581 volume:208 year:2021 pages:28-44 extent:17 https://doi.org/10.1016/j.biosystemseng.2021.05.006 Volltext GBV_USEFLAG_U GBV_ELV SYSFLAG_U SSG-OLC-PHA GBV_ILN_20 GBV_ILN_22 GBV_ILN_24 GBV_ILN_30 GBV_ILN_40 GBV_ILN_65 GBV_ILN_70 AR 208 2021 28-44 17 |
allfieldsSound |
10.1016/j.biosystemseng.2021.05.006 doi /cbs_pica/cbs_olc/import_discovery/elsevier/einzuspielen/GBV00000000001447.pica (DE-627)ELV054593840 (ELSEVIER)S1537-5110(21)00106-9 DE-627 ger DE-627 rakwb eng 630 VZ 640 VZ 320 VZ 630 640 610 VZ Rustia, Dan Jeric Arcega verfasserin aut Online semi-supervised learning applied to an automated insect pest monitoring system 2021transfer abstract 17 nicht spezifiziert zzz rdacontent nicht spezifiziert z rdamedia nicht spezifiziert zu rdacarrier The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition Elsevier Lu, Chen-Yi oth Chao, Jun-Jee oth Wu, Ya-Fang oth Chung, Jui-Yung oth Hsu, Ju-Chun oth Lin, Ta-Te oth Enthalten in Academ. Press Raggi, M. ELSEVIER The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) 2015 San Diego, Calif (DE-627)ELV018374581 volume:208 year:2021 pages:28-44 extent:17 https://doi.org/10.1016/j.biosystemseng.2021.05.006 Volltext GBV_USEFLAG_U GBV_ELV SYSFLAG_U SSG-OLC-PHA GBV_ILN_20 GBV_ILN_22 GBV_ILN_24 GBV_ILN_30 GBV_ILN_40 GBV_ILN_65 GBV_ILN_70 AR 208 2021 28-44 17 |
language |
English |
source |
Enthalten in The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) San Diego, Calif volume:208 year:2021 pages:28-44 extent:17 |
sourceStr |
Enthalten in The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) San Diego, Calif volume:208 year:2021 pages:28-44 extent:17 |
format_phy_str_mv |
Article |
institution |
findex.gbv.de |
topic_facet |
insect monitoring feature extraction integrated pest management semi-supervised learning image recognition |
dewey-raw |
630 |
isfreeaccess_bool |
false |
container_title |
The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) |
authorswithroles_txt_mv |
Rustia, Dan Jeric Arcega @@aut@@ Lu, Chen-Yi @@oth@@ Chao, Jun-Jee @@oth@@ Wu, Ya-Fang @@oth@@ Chung, Jui-Yung @@oth@@ Hsu, Ju-Chun @@oth@@ Lin, Ta-Te @@oth@@ |
publishDateDaySort_date |
2021-01-01T00:00:00Z |
hierarchy_top_id |
ELV018374581 |
dewey-sort |
3630 |
id |
ELV054593840 |
language_de |
englisch |
fullrecord |
<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01000caa a22002652 4500</leader><controlfield tag="001">ELV054593840</controlfield><controlfield tag="003">DE-627</controlfield><controlfield tag="005">20230626040455.0</controlfield><controlfield tag="007">cr uuu---uuuuu</controlfield><controlfield tag="008">210910s2021 xx |||||o 00| ||eng c</controlfield><datafield tag="024" ind1="7" ind2=" "><subfield code="a">10.1016/j.biosystemseng.2021.05.006</subfield><subfield code="2">doi</subfield></datafield><datafield tag="028" ind1="5" ind2="2"><subfield code="a">/cbs_pica/cbs_olc/import_discovery/elsevier/einzuspielen/GBV00000000001447.pica</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-627)ELV054593840</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(ELSEVIER)S1537-5110(21)00106-9</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-627</subfield><subfield code="b">ger</subfield><subfield code="c">DE-627</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1=" " ind2=" "><subfield code="a">eng</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">630</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">640</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">320</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">630</subfield><subfield code="a">640</subfield><subfield code="a">610</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="100" ind1="1" ind2=" "><subfield code="a">Rustia, Dan Jeric Arcega</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Online semi-supervised learning applied to an automated insect pest monitoring system</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="c">2021transfer abstract</subfield></datafield><datafield tag="300" ind1=" " ind2=" "><subfield code="a">17</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">nicht spezifiziert</subfield><subfield code="b">zzz</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">nicht spezifiziert</subfield><subfield code="b">z</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">nicht spezifiziert</subfield><subfield code="b">zu</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update.</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update.</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">insect monitoring</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">feature extraction</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">integrated pest management</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">semi-supervised learning</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">image recognition</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Lu, Chen-Yi</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Chao, Jun-Jee</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Wu, Ya-Fang</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Chung, Jui-Yung</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Hsu, Ju-Chun</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Lin, Ta-Te</subfield><subfield code="4">oth</subfield></datafield><datafield tag="773" ind1="0" ind2="8"><subfield code="i">Enthalten in</subfield><subfield code="n">Academ. Press</subfield><subfield code="a">Raggi, M. ELSEVIER</subfield><subfield code="t">The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy)</subfield><subfield code="d">2015</subfield><subfield code="g">San Diego, Calif</subfield><subfield code="w">(DE-627)ELV018374581</subfield></datafield><datafield tag="773" ind1="1" ind2="8"><subfield code="g">volume:208</subfield><subfield code="g">year:2021</subfield><subfield code="g">pages:28-44</subfield><subfield code="g">extent:17</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doi.org/10.1016/j.biosystemseng.2021.05.006</subfield><subfield code="3">Volltext</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_USEFLAG_U</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ELV</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SYSFLAG_U</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SSG-OLC-PHA</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_20</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_22</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_24</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_30</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_40</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_65</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_70</subfield></datafield><datafield tag="951" ind1=" " ind2=" "><subfield code="a">AR</subfield></datafield><datafield tag="952" ind1=" " ind2=" "><subfield code="d">208</subfield><subfield code="j">2021</subfield><subfield code="h">28-44</subfield><subfield code="g">17</subfield></datafield></record></collection>
|
author |
Rustia, Dan Jeric Arcega |
spellingShingle |
Rustia, Dan Jeric Arcega ddc 630 ddc 640 ddc 320 Elsevier insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition Online semi-supervised learning applied to an automated insect pest monitoring system |
authorStr |
Rustia, Dan Jeric Arcega |
ppnlink_with_tag_str_mv |
@@773@@(DE-627)ELV018374581 |
format |
electronic Article |
dewey-ones |
630 - Agriculture & related technologies 640 - Home & family management 320 - Political science 610 - Medicine & health |
delete_txt_mv |
keep |
author_role |
aut |
collection |
elsevier |
remote_str |
true |
illustrated |
Not Illustrated |
topic_title |
630 VZ 640 VZ 320 VZ 630 640 610 VZ Online semi-supervised learning applied to an automated insect pest monitoring system insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition Elsevier |
topic |
ddc 630 ddc 640 ddc 320 Elsevier insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition |
topic_unstemmed |
ddc 630 ddc 640 ddc 320 Elsevier insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition |
topic_browse |
ddc 630 ddc 640 ddc 320 Elsevier insect monitoring Elsevier feature extraction Elsevier integrated pest management Elsevier semi-supervised learning Elsevier image recognition |
format_facet |
Elektronische Aufsätze Aufsätze Elektronische Ressource |
format_main_str_mv |
Text Zeitschrift/Artikel |
carriertype_str_mv |
zu |
author2_variant |
c y l cyl j j c jjc y f w yfw j y c jyc j c h jch t t l ttl |
hierarchy_parent_title |
The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) |
hierarchy_parent_id |
ELV018374581 |
dewey-tens |
630 - Agriculture 640 - Home & family management 320 - Political science 610 - Medicine & health |
hierarchy_top_title |
The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) |
isfreeaccess_txt |
false |
familylinks_str_mv |
(DE-627)ELV018374581 |
title |
Online semi-supervised learning applied to an automated insect pest monitoring system |
ctrlnum |
(DE-627)ELV054593840 (ELSEVIER)S1537-5110(21)00106-9 |
title_full |
Online semi-supervised learning applied to an automated insect pest monitoring system |
author_sort |
Rustia, Dan Jeric Arcega |
journal |
The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) |
journalStr |
The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy) |
lang_code |
eng |
isOA_bool |
false |
dewey-hundreds |
600 - Technology 300 - Social sciences |
recordtype |
marc |
publishDateSort |
2021 |
contenttype_str_mv |
zzz |
container_start_page |
28 |
author_browse |
Rustia, Dan Jeric Arcega |
container_volume |
208 |
physical |
17 |
class |
630 VZ 640 VZ 320 VZ 630 640 610 VZ |
format_se |
Elektronische Aufsätze |
author-letter |
Rustia, Dan Jeric Arcega |
doi_str_mv |
10.1016/j.biosystemseng.2021.05.006 |
dewey-full |
630 640 320 610 |
title_sort |
online semi-supervised learning applied to an automated insect pest monitoring system |
title_auth |
Online semi-supervised learning applied to an automated insect pest monitoring system |
abstract |
The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. |
abstractGer |
The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. |
abstract_unstemmed |
The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update. |
collection_details |
GBV_USEFLAG_U GBV_ELV SYSFLAG_U SSG-OLC-PHA GBV_ILN_20 GBV_ILN_22 GBV_ILN_24 GBV_ILN_30 GBV_ILN_40 GBV_ILN_65 GBV_ILN_70 |
title_short |
Online semi-supervised learning applied to an automated insect pest monitoring system |
url |
https://doi.org/10.1016/j.biosystemseng.2021.05.006 |
remote_bool |
true |
author2 |
Lu, Chen-Yi Chao, Jun-Jee Wu, Ya-Fang Chung, Jui-Yung Hsu, Ju-Chun Lin, Ta-Te |
author2Str |
Lu, Chen-Yi Chao, Jun-Jee Wu, Ya-Fang Chung, Jui-Yung Hsu, Ju-Chun Lin, Ta-Te |
ppnlink |
ELV018374581 |
mediatype_str_mv |
z |
isOA_txt |
false |
hochschulschrift_bool |
false |
author2_role |
oth oth oth oth oth oth |
doi_str |
10.1016/j.biosystemseng.2021.05.006 |
up_date |
2024-07-06T22:09:28.347Z |
_version_ |
1803869244200321024 |
fullrecord_marcxml |
<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01000caa a22002652 4500</leader><controlfield tag="001">ELV054593840</controlfield><controlfield tag="003">DE-627</controlfield><controlfield tag="005">20230626040455.0</controlfield><controlfield tag="007">cr uuu---uuuuu</controlfield><controlfield tag="008">210910s2021 xx |||||o 00| ||eng c</controlfield><datafield tag="024" ind1="7" ind2=" "><subfield code="a">10.1016/j.biosystemseng.2021.05.006</subfield><subfield code="2">doi</subfield></datafield><datafield tag="028" ind1="5" ind2="2"><subfield code="a">/cbs_pica/cbs_olc/import_discovery/elsevier/einzuspielen/GBV00000000001447.pica</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-627)ELV054593840</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(ELSEVIER)S1537-5110(21)00106-9</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-627</subfield><subfield code="b">ger</subfield><subfield code="c">DE-627</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1=" " ind2=" "><subfield code="a">eng</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">630</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">640</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">320</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">630</subfield><subfield code="a">640</subfield><subfield code="a">610</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="100" ind1="1" ind2=" "><subfield code="a">Rustia, Dan Jeric Arcega</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Online semi-supervised learning applied to an automated insect pest monitoring system</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="c">2021transfer abstract</subfield></datafield><datafield tag="300" ind1=" " ind2=" "><subfield code="a">17</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">nicht spezifiziert</subfield><subfield code="b">zzz</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">nicht spezifiziert</subfield><subfield code="b">z</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">nicht spezifiziert</subfield><subfield code="b">zu</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update.</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">The unavailability and variability of training samples are the two essential concerns in the training of deep neural network models for image classification. For automated image monitoring systems, these problems are difficult when training a model through supervised learning methods because of the time and effort required. This paper proposes an adaptive solution to this problem by applying online semi-supervised learning to an automated insect pest monitoring system. The method used includes unsupervised pseudo-labelling of insect images and the training of semi-supervised classifier models for insect image recognition. The pseudo-labelling algorithm includes three major components: image labelling, label reconfirmation, and sample cleaning. Experiments were conducted on two unlabelled 1-year insect image datasets to evaluate the efficacy of the proposed method. It was found that the pseudo-labelling algorithm could achieve accuracy up to 0.963, hence enabling automated training data collection. The temporal improvement of the insect recognition performance by including new training data to retrain the classifier model was comparable in performance to the supervised learning approach as evaluated by cluster density, silhouette score, and F 1-score. The proposed method was also able to automatically collect quality samples and train models regardless of the complexity of the images, making it a good alternative to replace laborious supervised learning. The proposed method can prevent contamination of a training dataset when images from new locations are collected. The presented techniques may also be used in other continuous learning applications that require automated training data collection and online model update.</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">insect monitoring</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">feature extraction</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">integrated pest management</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">semi-supervised learning</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">image recognition</subfield><subfield code="2">Elsevier</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Lu, Chen-Yi</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Chao, Jun-Jee</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Wu, Ya-Fang</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Chung, Jui-Yung</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Hsu, Ju-Chun</subfield><subfield code="4">oth</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Lin, Ta-Te</subfield><subfield code="4">oth</subfield></datafield><datafield tag="773" ind1="0" ind2="8"><subfield code="i">Enthalten in</subfield><subfield code="n">Academ. Press</subfield><subfield code="a">Raggi, M. ELSEVIER</subfield><subfield code="t">The role of policy priorities and targeting in the spatial location of participation in Agri-Environmental Schemes in Emilia-Romagna (Italy)</subfield><subfield code="d">2015</subfield><subfield code="g">San Diego, Calif</subfield><subfield code="w">(DE-627)ELV018374581</subfield></datafield><datafield tag="773" ind1="1" ind2="8"><subfield code="g">volume:208</subfield><subfield code="g">year:2021</subfield><subfield code="g">pages:28-44</subfield><subfield code="g">extent:17</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doi.org/10.1016/j.biosystemseng.2021.05.006</subfield><subfield code="3">Volltext</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_USEFLAG_U</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ELV</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SYSFLAG_U</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SSG-OLC-PHA</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_20</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_22</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_24</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_30</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_40</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_65</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_70</subfield></datafield><datafield tag="951" ind1=" " ind2=" "><subfield code="a">AR</subfield></datafield><datafield tag="952" ind1=" " ind2=" "><subfield code="d">208</subfield><subfield code="j">2021</subfield><subfield code="h">28-44</subfield><subfield code="g">17</subfield></datafield></record></collection>
|
score |
7.399932 |