An automatic method for removing empty camera trap images using ensemble learning
Abstract Camera traps often produce massive images, and empty images that do not contain animals are usually overwhelming. Deep learning is a machine‐learning algorithm and widely used to identify empty camera trap images automatically. Existing methods with high accuracy are based on millions of tr...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2021-06-01
|
| Series: | Ecology and Evolution |
| Subjects: | |
| Online Access: | https://doi.org/10.1002/ece3.7591 |
| _version_ | 1818438235529936896 |
|---|---|
| author | Deng‐Qi Yang Kun Tan Zhi‐Pang Huang Xiao‐Wei Li Ben‐Hui Chen Guo‐Peng Ren Wen Xiao |
| author_facet | Deng‐Qi Yang Kun Tan Zhi‐Pang Huang Xiao‐Wei Li Ben‐Hui Chen Guo‐Peng Ren Wen Xiao |
| author_sort | Deng‐Qi Yang |
| collection | DOAJ |
| description | Abstract Camera traps often produce massive images, and empty images that do not contain animals are usually overwhelming. Deep learning is a machine‐learning algorithm and widely used to identify empty camera trap images automatically. Existing methods with high accuracy are based on millions of training samples (images) and require a lot of time and personnel costs to label the training samples manually. Reducing the number of training samples can save the cost of manually labeling images. However, the deep learning models based on a small dataset produce a large omission error of animal images that many animal images tend to be identified as empty images, which may lead to loss of the opportunities of discovering and observing species. Therefore, it is still a challenge to build the DCNN model with small errors on a small dataset. Using deep convolutional neural networks and a small‐size dataset, we proposed an ensemble learning approach based on conservative strategies to identify and remove empty images automatically. Furthermore, we proposed three automatic identifying schemes of empty images for users who accept different omission errors of animal images. Our experimental results showed that these three schemes automatically identified and removed 50.78%, 58.48%, and 77.51% of the empty images in the dataset when the omission errors were 0.70%, 1.13%, and 2.54%, respectively. The analysis showed that using our scheme to automatically identify empty images did not omit species information. It only slightly changed the frequency of species occurrence. When only a small dataset was available, our approach provided an alternative to users to automatically identify and remove empty images, which can significantly reduce the time and personnel costs required to manually remove empty images. The cost savings were comparable to the percentage of empty images removed by models. |
| first_indexed | 2024-12-14T17:37:21Z |
| format | Article |
| id | doaj.art-df29b24ad22f4133a5ca90304c5a2628 |
| institution | Directory Open Access Journal |
| issn | 2045-7758 |
| language | English |
| last_indexed | 2024-12-14T17:37:21Z |
| publishDate | 2021-06-01 |
| publisher | Wiley |
| record_format | Article |
| series | Ecology and Evolution |
| spelling | doaj.art-df29b24ad22f4133a5ca90304c5a26282022-12-21T22:52:56ZengWileyEcology and Evolution2045-77582021-06-0111127591760110.1002/ece3.7591An automatic method for removing empty camera trap images using ensemble learningDeng‐Qi Yang0Kun Tan1Zhi‐Pang Huang2Xiao‐Wei Li3Ben‐Hui Chen4Guo‐Peng Ren5Wen Xiao6Department of Mathematics and Computer Science Dali University Dali ChinaInstitute of Eastern‐Himalaya Biodiversity Research Dali University Dali ChinaInstitute of Eastern‐Himalaya Biodiversity Research Dali University Dali ChinaDepartment of Mathematics and Computer Science Dali University Dali ChinaDepartment of Mathematics and Computer Science Dali University Dali ChinaInstitute of Eastern‐Himalaya Biodiversity Research Dali University Dali ChinaInstitute of Eastern‐Himalaya Biodiversity Research Dali University Dali ChinaAbstract Camera traps often produce massive images, and empty images that do not contain animals are usually overwhelming. Deep learning is a machine‐learning algorithm and widely used to identify empty camera trap images automatically. Existing methods with high accuracy are based on millions of training samples (images) and require a lot of time and personnel costs to label the training samples manually. Reducing the number of training samples can save the cost of manually labeling images. However, the deep learning models based on a small dataset produce a large omission error of animal images that many animal images tend to be identified as empty images, which may lead to loss of the opportunities of discovering and observing species. Therefore, it is still a challenge to build the DCNN model with small errors on a small dataset. Using deep convolutional neural networks and a small‐size dataset, we proposed an ensemble learning approach based on conservative strategies to identify and remove empty images automatically. Furthermore, we proposed three automatic identifying schemes of empty images for users who accept different omission errors of animal images. Our experimental results showed that these three schemes automatically identified and removed 50.78%, 58.48%, and 77.51% of the empty images in the dataset when the omission errors were 0.70%, 1.13%, and 2.54%, respectively. The analysis showed that using our scheme to automatically identify empty images did not omit species information. It only slightly changed the frequency of species occurrence. When only a small dataset was available, our approach provided an alternative to users to automatically identify and remove empty images, which can significantly reduce the time and personnel costs required to manually remove empty images. The cost savings were comparable to the percentage of empty images removed by models.https://doi.org/10.1002/ece3.7591artificial intelligencecamera trap imagesconvolutional neural networksdeep learningensemble learning |
| spellingShingle | Deng‐Qi Yang Kun Tan Zhi‐Pang Huang Xiao‐Wei Li Ben‐Hui Chen Guo‐Peng Ren Wen Xiao An automatic method for removing empty camera trap images using ensemble learning Ecology and Evolution artificial intelligence camera trap images convolutional neural networks deep learning ensemble learning |
| title | An automatic method for removing empty camera trap images using ensemble learning |
| title_full | An automatic method for removing empty camera trap images using ensemble learning |
| title_fullStr | An automatic method for removing empty camera trap images using ensemble learning |
| title_full_unstemmed | An automatic method for removing empty camera trap images using ensemble learning |
| title_short | An automatic method for removing empty camera trap images using ensemble learning |
| title_sort | automatic method for removing empty camera trap images using ensemble learning |
| topic | artificial intelligence camera trap images convolutional neural networks deep learning ensemble learning |
| url | https://doi.org/10.1002/ece3.7591 |
| work_keys_str_mv | AT dengqiyang anautomaticmethodforremovingemptycameratrapimagesusingensemblelearning AT kuntan anautomaticmethodforremovingemptycameratrapimagesusingensemblelearning AT zhipanghuang anautomaticmethodforremovingemptycameratrapimagesusingensemblelearning AT xiaoweili anautomaticmethodforremovingemptycameratrapimagesusingensemblelearning AT benhuichen anautomaticmethodforremovingemptycameratrapimagesusingensemblelearning AT guopengren anautomaticmethodforremovingemptycameratrapimagesusingensemblelearning AT wenxiao anautomaticmethodforremovingemptycameratrapimagesusingensemblelearning AT dengqiyang automaticmethodforremovingemptycameratrapimagesusingensemblelearning AT kuntan automaticmethodforremovingemptycameratrapimagesusingensemblelearning AT zhipanghuang automaticmethodforremovingemptycameratrapimagesusingensemblelearning AT xiaoweili automaticmethodforremovingemptycameratrapimagesusingensemblelearning AT benhuichen automaticmethodforremovingemptycameratrapimagesusingensemblelearning AT guopengren automaticmethodforremovingemptycameratrapimagesusingensemblelearning AT wenxiao automaticmethodforremovingemptycameratrapimagesusingensemblelearning |