<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>The behaviours of the pig are often closely related to their health. Pig recognition is very important for pig behaviour analysis and digital breeding. Currently, the early signs and abnormal behaviours of sick pigs in breeding farms are mainly completed by human observation. However, visual inspection is labour intensive and time-consuming, and it suffers from the problems of individual experiences and varying environments. An improved ResNet model was proposed and applied to detect individual pigs in this study based on deep learning knowledge. The developed model captured the features of pigs applying across layer connections, and the ability of feature expression was improved by adding a new residual module. The number of layers was reduced to minimise the net complexity. Generally, the ResNet frame was developed by reducing the number of convolution layers, constructing different types of the residual module and adding the number of convolution kernels. The training accuracy and testing accuracy reached 98.2% and 96.4%, respectively, when using the improved model. The experiment results showed that the method proposed in this paper for checking living situations and disease prevention of commercial pigs in pig farms is potential.</p>
</abstract>

The behaviours of the pig are often closely related to their health. Pig recognition is very important for pig behaviour analysis and digital breeding. Currently, the early signs and abnormal behaviours of sick pigs in breeding farms are mainly completed by human observation. However, visual inspection is labour intensive and time-consuming, and it suffers from the problems of individual experiences and varying environments. An improved ResNet model was proposed and applied to detect individual pigs in this study based on deep learning knowledge. The developed model captured the features of pigs applying across layer connections, and the ability of feature expression was improved by adding a new residual module. The number of layers was reduced to minimise the net complexity. Generally, the ResNet frame was developed by reducing the number of convolution layers, constructing different types of the residual module and adding the number of convolution kernels. The training accuracy and testing accuracy reached 98.2% and 96.4%, respectively, when using the improved model. The experiment results showed that the method proposed in this paper for checking living situations and disease prevention of commercial pigs in pig farms is potential.

Detection of pig based on improved RESNET model in natural scene

Applied Mathematics and Nonlinear Sciences

This work is licensed under the Creative Commons Attribution 4.0 International License.

{"article-title":"Detection of pig based on improved RESNET model in natural scene"}

The behaviours of the pig are often closely related to their health. Pig recognition is very important for pig behaviour analysis and digital breeding....

CNN model	Training accuracy (%)	Testing accuracy (%)	Loss (%)
AlexNet	94.2	90.4	22.1
GoogLeNet	96.7	93.1	16.7
ResNet	97.7	95.3	10.7
Improved ResNet	98.2	96.4	7.7

Pig posture	The number of original images	The number of images after expanding	The images in the training set	The images in the testing set
Standing	600	1200	983	217
Standing with lowering the head	200	400	316	84
Raising head	240	480	390	90
Laying down	55	110	88	22

Pig posture	The number of original images	The number of images after expanding	The images in the training set	The images in the testing set
Standing	600	4800	3844	956
Standing with lowering the head	200	1600	1264	336
Raising head	240	1920	1560	360
Laying down	55	440	352	88

Detection of pig based on improved RESNET model in natural scene

Online veröffentlicht: 30. Okt. 2021

Seitenbereich: 215 - 226

Eingereicht: 26. März 2021

Akzeptiert: 29. Juni 2021

DOI: https://doi.org/10.2478/amns.2021.2.00040

Schlüsselwörter
deep learning, object detection, ResNet model, feature classification

© 2021 Weixian Song et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Fig. 7

Fig. 8

Fig. 9

Fig. 10

Fig. 11

Fig. 12

Fig. 13

Training parameter settings of this study.

The training and testing results of dataset2 of different CNN models.

Distribution of training and testing images for dataset 1.

Distribution of training and testing images for dataset 2.

Detection of pig based on improved RESNET model in natural scene

Online veröffentlicht: 30. Okt. 2021

Seitenbereich: 215 - 226

Eingereicht: 26. März 2021

Akzeptiert: 29. Juni 2021

DOI: https://doi.org/10.2478/amns.2021.2.00040

Schlüsselwörterdeep learning, object detection, ResNet model, feature classification

© 2021 Weixian Song et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Fig. 7

Fig. 8

Fig. 9

Fig. 10

Fig. 11

Fig. 12

Fig. 13

Training parameter settings of this study.

The training and testing results of dataset2 of different CNN models.

Distribution of training and testing images for dataset 1.

Distribution of training and testing images for dataset 2.

Schlüsselwörter
deep learning, object detection, ResNet model, feature classification