Enhanced Mask R-CNN for herd segmentation

Abstract

Livestock image segmentation is an important task in the field of vision and image processing. Since utilizing the concentration of forage in the grazing area with shielding the surrounding farm plants and crops is necessary for making effective cattle ranch arrangements, there is a need for a segmentation method that can handle multiple objects segmentation. Moreover, the indistinct boundaries and irregular shapes of cattle bodies discourage the application of the existing Mask Region-based Convolutional Neural Network (Mask R-CNN) which was primarily modeled for the segmentation of natural images. To address this, an enhanced Mask R-CNN model was proposed for multiple objects instance segmentation to support indistinct boundaries and irregular shapes of cattle bodies for precision livestock farming. The contributions of this method are in multiple folds: 1) optimal filter size smaller than a residual network for extracting smaller and composite features; 2) region proposals for utilizing multiscale semantic features; 3) Mask R-CNN's fully connected layer integrated with sub-network for an enhanced segmentation. The experiment conducted on pre-processed datasets produced a mean average precision (mAP) of 0.93, which was higher than the results from the existing state-of-the-art models.

Application of Machine Learning Identification and Classification of Muturu and Keteku Cattle Species for a Smart Agricultural Practice in Developing Countries such as Nigeria

“…For the body patterns identification, neural networks with convolutional layers were used. Classification of individual cattle using instance segmentation was by using enhanced mask R-CNN (Bello et al, 2021b). The essence of these approaches is to mitigate the challenges of using the traditional cattle identification and classification methods which most of the time cause injury and even the death of the animal.…”

“…Based on these assertions, this paper proposed mainstream model that enables precision and intelligence perception for a smart agricultural practice. The widely acceptability of convolution neural network-based methods with ability to achieve accurate result-oriented image feature extraction and representation has continue to grow in the field of vision and image processing (Bello et al, 2021a;Bello et al, 2021b;Bello et al, 2021c). Although, these methods have been widely applied for animal identification purposes such as cattle without the need for pre-specifying any features (Kumar et al, 2018) but, the application of CNN-based methods in monitoring the activities of animals in animal farming is still not widespread (Li et al, 2021).…”

A real-time and accurate detection approach for bucket teeth falling off based on improved YOLOX

“…The spatial pyramid pooling (SPP) layer performs feature extraction on the maximum pooling of different pooling kernel sizes, improving the receptive field of the network, and enhancing the nonlinear expression ability of the network. Neck improves the feature pyramid network + path aggregation network (FPN + PAN) structure to enhance the feature fusion of the network and the transmission of inference information in the network (Bello et al, 2021). Prediction replaces the original classification and regression layer couple head with decouple head, adopting anchor-free frame mechanism and simOTA label assignment method to solve the optimal transmission problem and reduce the amount of model parameters, which significantly improves the regression speed and accuracy of the network.…”

Underwater Image Enhancement for Instance Segmentation using Deep Learning Models

“…Fig.2. Methodology for underwater instance segmentation using Mask R-CNN(Bello et al, 2021a;Bello et al, 2021b) …”