Hatching egg classification based on CNN with channel weighting and joint supervision

Abstract

Convolutional neural networks (CNNs) show state-of-the-art performance in tackling a variety of visual tasks. It is expected that a CNN can be applied to the 9-day hatching eggs classification. These hatching eggs are divided into fertile eggs and dead eggs. Because of the inter-class similarity and intra-class difference issues in 9-day hatching eggs datasets, the CNN classification method combining channel weighting (squeeze-and-excitation module) and joint supervision is proposed to improve the classification accuracy. We use the center loss and softmax loss together as a joint supervision signal. With such joint supervision, the CNN can obtain the deep features with inter-class dispersion and intra-class compactness, which enhances the discriminative and generalization powers. Simultaneously, channel weighting is adopted in feature extraction, which is added in each convolutional layer to make better use of the channel features. The experimental results demonstrate that the proposed method successfully solves the classification problem of hatching eggs. The accuracy of our method is 98.8%.