Attention mechanism CNN-based 5-day and 9-day incubated egg embryo image classification method

Abstract

The invention relates to an attention mechanism CNN-based 5-day and 9-day incubated egg embryo image classification method so as to realize the egg embryo activity detection and sorting. The method comprises the following steps of 1) proposing to use the depth separable convolution to realize the more sufficient feature extraction on an existing channel attention feature map; 2) generating a high-resolution and large-receptive field attention weight map by adopting the spatial pyramid cavity convolution; and 3) performing the spatial weighting on the feature map by the weight mask through element-by-element multiplication, so that the effect of enhancing the useful information and suppressing the noise can be realized. Results show that the attention module plays a role of a feature selector in the convolutional neural network, and the feature expression capability of the convolutional neural network is enhanced, so that the classification accuracy is improved, and the problems of lowefficiency, high labor consumption and the like of manually sorting the 5-day and 9-day incubated egg embryos are successfully solved.

Description

technical field [0001] The invention relates to a method for classifying images of 5-day and 9-day hatched egg embryos based on the attention mechanism CNN. Some technologies are more stable and have good classification performance, and belong to the field of biological image recognition and deep learning computer vision. Background technique [0002] The prevention of avian influenza is mainly through vaccination. At present, the preparation of avian influenza vaccine is mainly carried out by inoculating the virus strain in egg embryos and then multiplying and culturing. Due to the difference in the inoculation position of the virus strain, non-specific death will occur in the embryo eggs of the virus strain during the culture process. During the cultivation of embryonated eggs of the virus strain, dead embryonated eggs that are not eliminated will lead to the failure of strain proliferation and culture. Therefore, the detection and classification of the viability of the e...

AI Technical Summary

Problems solved by technology

However, due to the characteristics of the above-mentioned 9-day embryos, weak embryos still have local sparse blood vessels, and the general classification network is easy to misjudge weak embryos as live embryos, which affects the final accuracy rate.

In order to solve the above problems, the present invention proposes a convolutional neural network with a novel attention mechanism, which can better guide the network to locate the most distinguishing features between learning classes, and can complete the detection of 5-day embryos and 9-day embryos. Classification with high accuracy and a larger discriminative receptive field, thus alleviating the problem of misjudgment of weak embryos with local sparse blood vessels

Images