Deep learning framework for bovine iris segmentation

Iris segmentation is an initial step for identifying the biometrics of animals when establishing a traceability system for livestock. In this study, we propose a deep learning framework for pixel-wise segmentation of bovine iris with a minimized use of annotation labels utilizing the BovineAAEyes80 public dataset. The proposed image segmentation framework encompasses data collection, data preparation, data augmentation selection, training of 15 deep neural network (DNN) models with varying encoder backbones and segmentation decoder DNNs, and evaluation of the models using multiple metrics and graphical segmentation results. This framework aims to provide comprehensive and in-depth information on each model's training and testing outcomes to optimize bovine iris segmentation performance. In the experiment, U-Net with a VGG16 backbone was identified as the optimal combination of encoder and decoder models for the dataset, achieving an accuracy and dice coefficient score of 99.50% and 98.35%, respectively. Notably, the selected model accurately segmented even corrupted images without proper annotation data. This study contributes to the advancement of iris segmentation and the establishment of a reliable DNN training framework.

Cattle face recognition under partial occlusion

The occlusion in the real feedlot environment is ubiquitous, and the current research based on the cattle face recognition under occlusion conditions is almost non-existent. Thus, an attention mechanism module with high accuracy and low model complexity is designed to incorporate into MobileNet so that the cattle face under occlusion can be identify accurately, which is the RGB images captured in the ranch environment. In this paper, we also construct a Simmental cattle face image dataset for data modeling and method evaluation, which contains 10,239 images of 103 cattle. The experimental results show that when the occluder is in the upper left and lower right corner, if the occlusion rate is less than 30%, the value of Top_1 reaches more than 90%; if it is less than 50%, the value of Top_1 is more than 80%. Even if the middle part occludes lots of important information, the occlusion rate of 40% has an accuracy of more than 80%. Furthermore, comparing the proposal model with MobileNet, the parameter and model size are equal, and the amount of calculation as a cost increase a little. Therefore, the proposal model is suitable to transplant to the embedded system in the future.

A Cascaded Model Based on EfficientDet and YOLACT++ for Instance Segmentation of Cow Collar ID Tag in an Image

In recent years, many imaging systems have been developed to monitor the physiological and behavioral status of dairy cows. However, most of these systems do not have the ability to identify individual cows because the systems need to cooperate with radio frequency identification (RFID) to collect information about individual animals. The distance at which RFID can identify a target is limited, and matching the identified targets in a scenario of multitarget images is difficult. To solve the above problems, we constructed a cascaded method based on cascaded deep learning models, to detect and segment a cow collar ID tag in an image. First, EfficientDet-D4 was used to detect the ID tag area of the image, and then, YOLACT++ was used to segment the area of the tag to realize the accurate segmentation of the ID tag when the collar area accounts for a small proportion of the image. In total, 938 and 406 images of cows with collar ID tags, which were collected at Coldstream Research Dairy Farm, University of Kentucky, USA, in August 2016, were used to train and test the two models, respectively. The results showed that the average precision of the EfficientDet-D4 model reached 96.5% when the intersection over union (IoU) was set to 0.5, and the average precision of the YOLACT++ model reached 100% when the IoU was set to 0.75. The overall accuracy of the cascaded model was 96.5%, and the processing time of a single frame image was 1.92 s. The performance of the cascaded model proposed in this paper is better than that of the common instance segmentation models, and it is robust to changes in brightness, deformation, and interference around the tag.