Deep Learning Pig Weight Estimation via Keypoint Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating the body size and weight of pigs are inefficient and labor-intensive, often stimulating the animals and failing to effectively handle nonlinear data, which limits their precision and accuracy.
Innovation Solution
A deep learning method using a Keypoint-RCNN algorithm for keypoint detection and a ResNext-101 feature extraction network for weight estimation, which includes instance segmentation and a softmax layer, to predict body size and weight without the need for feature engineering, improving handling of noisy and nonlinear data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional manual measurement methods are used, then measurement can be performed, but it requires a lot of time and manpower and is inefficient
Solution Approach 1:
The patent replaces manual mechanical measurement with an automated computer vision system using deep learning algorithms. The system captures images of pigs and automatically estimates body size and weight through neural network models, eliminating the need for manual measurement operations and significantly improving measurement efficiency while reducing time consumption.
2Object-affected harmful factors
If traditional manual measurement methods are used, then measurement can be performed, but it is easy to stimulate pig bodies which is not conducive to pig welfare
Solution Approach 1:
The patent replaces physical contact-based manual measurement with non-contact image-based measurement. By using cameras and deep learning algorithms to estimate pig body dimensions from images, the system eliminates physical contact that would stimulate or stress the animals, thereby improving pig welfare while maintaining measurement capability.
3Measurement precision
If traditional linear regression models are used for weight estimation, then the model is simple, but it is inferior to deep learning methods in processing noisy data and dealing with nonlinear problems
Solution Approach 1:
The patent transitions from simple linear regression models to complex deep learning models (including ResNet and U-Net architectures). This parameter change in model complexity enables the system to process noisy image data and capture nonlinear relationships between image features and pig weight, significantly improving measurement precision despite the increased model complexity.
Data Source
AI summary
The present disclosure provides a method for estimating a body size and weight of a pig based on deep learning, and relates to the technical field of deep learning. The present disclosure predicts the weight of the pig by using a convolutional neural network. Relevant features are learned by the convolutional neural network, and feature engineering extraction is not needed to be established, so that extracted features are more comprehensive, and the convolutional neural network is superior to a linear model in processing of noisy data and nonlinear problems of data. Images of the pig are shot by an ordinary two-dimensional (2d) color camera.