Cattle Weight Estimation Using Fully and Weakly Supervised Segmentation from 2D Images
Weight information is important in cattle breeding because it can measure animal growth and be used to calculate the appropriate amount of daily feed. To estimate the weight, we developed an image-based method that does not stress cattle and requires no manual labor. From a 2D image, a mask was obta...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-02-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/5/2896 |
| _version_ | 1827753098058137600 |
|---|---|
| author | Chang-bok Lee Han-sung Lee Hyun-chong Cho |
| author_facet | Chang-bok Lee Han-sung Lee Hyun-chong Cho |
| author_sort | Chang-bok Lee |
| collection | DOAJ |
| description | Weight information is important in cattle breeding because it can measure animal growth and be used to calculate the appropriate amount of daily feed. To estimate the weight, we developed an image-based method that does not stress cattle and requires no manual labor. From a 2D image, a mask was obtained by segmenting the animal and background, and weights were estimated using a deep neural network with residual connections by extracting weight-related features from the segmentation mask. Two image segmentation methods, fully and weakly supervised segmentation, were compared. The fully supervised segmentation method uses a Mask R-CNN model that learns the ground truth mask generated by labeling as the correct answer. The weakly supervised segmentation method uses an activation visualization map that is proposed in this study. The first method creates a more precise mask, but the second method does not require ground truth segmentation labeling. The body weight was estimated using statistical features of the segmented region. In experiments, the following performance results were obtained: a mean average error of 17.31 kg and mean absolute percentage error of 5.52% for fully supervised segmentation, and a mean average error of 35.91 kg and mean absolute percentage error of 10.1% for the weakly supervised segmentation. |
| first_indexed | 2024-03-11T07:32:22Z |
| format | Article |
| id | doaj.art-ea12399193934e33911d3da63ee42e2e |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-11T07:32:22Z |
| publishDate | 2023-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-ea12399193934e33911d3da63ee42e2e2023-11-17T07:16:31ZengMDPI AGApplied Sciences2076-34172023-02-01135289610.3390/app13052896Cattle Weight Estimation Using Fully and Weakly Supervised Segmentation from 2D ImagesChang-bok Lee0Han-sung Lee1Hyun-chong Cho2Department Graduate Program for BIT Medical Convergence, Kangwon National University, Chuncheon 24341, Republic of KoreaDepartment Graduate Program for BIT Medical Convergence, Kangwon National University, Chuncheon 24341, Republic of KoreaDepartment Graduate Program for BIT Medical Convergence, Kangwon National University, Chuncheon 24341, Republic of KoreaWeight information is important in cattle breeding because it can measure animal growth and be used to calculate the appropriate amount of daily feed. To estimate the weight, we developed an image-based method that does not stress cattle and requires no manual labor. From a 2D image, a mask was obtained by segmenting the animal and background, and weights were estimated using a deep neural network with residual connections by extracting weight-related features from the segmentation mask. Two image segmentation methods, fully and weakly supervised segmentation, were compared. The fully supervised segmentation method uses a Mask R-CNN model that learns the ground truth mask generated by labeling as the correct answer. The weakly supervised segmentation method uses an activation visualization map that is proposed in this study. The first method creates a more precise mask, but the second method does not require ground truth segmentation labeling. The body weight was estimated using statistical features of the segmented region. In experiments, the following performance results were obtained: a mean average error of 17.31 kg and mean absolute percentage error of 5.52% for fully supervised segmentation, and a mean average error of 35.91 kg and mean absolute percentage error of 10.1% for the weakly supervised segmentation.https://www.mdpi.com/2076-3417/13/5/2896convolutional neural networkdeep learningfully supervised segmentationweakly supervised segmentationcattle weight estimation |
| spellingShingle | Chang-bok Lee Han-sung Lee Hyun-chong Cho Cattle Weight Estimation Using Fully and Weakly Supervised Segmentation from 2D Images Applied Sciences convolutional neural network deep learning fully supervised segmentation weakly supervised segmentation cattle weight estimation |
| title | Cattle Weight Estimation Using Fully and Weakly Supervised Segmentation from 2D Images |
| title_full | Cattle Weight Estimation Using Fully and Weakly Supervised Segmentation from 2D Images |
| title_fullStr | Cattle Weight Estimation Using Fully and Weakly Supervised Segmentation from 2D Images |
| title_full_unstemmed | Cattle Weight Estimation Using Fully and Weakly Supervised Segmentation from 2D Images |
| title_short | Cattle Weight Estimation Using Fully and Weakly Supervised Segmentation from 2D Images |
| title_sort | cattle weight estimation using fully and weakly supervised segmentation from 2d images |
| topic | convolutional neural network deep learning fully supervised segmentation weakly supervised segmentation cattle weight estimation |
| url | https://www.mdpi.com/2076-3417/13/5/2896 |
| work_keys_str_mv | AT changboklee cattleweightestimationusingfullyandweaklysupervisedsegmentationfrom2dimages AT hansunglee cattleweightestimationusingfullyandweaklysupervisedsegmentationfrom2dimages AT hyunchongcho cattleweightestimationusingfullyandweaklysupervisedsegmentationfrom2dimages |