A method of generating image recognition model

Abstract

The invention discloses a method for generating an image recognition model. The method comprises the steps of obtaining a training image with annotation data; inputting the training image belonging tothe target category into a pre-trained main network and a pre-trained first branch network for processing to output feature point coordinates of the target object, and training to obtain an intermediate main network and an intermediate first branch network on the basis of the annotation data and the output feature point coordinates; inputting the training images belonging to the target category and not belonging to the target category into an intermediate main network and a pre-trained second branch network for processing to output the category of the training images, and training to obtain the main network and the second branch network based on the annotation data and the output category; and generating an image recognition model based on the trained main network, the first branch network and the second branch network. According to the scheme, the detection precision and stability of the target object feature points in the image can be improved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method for generating an image recognition model, an image recognition method, a computing device and a storage medium. Background technique [0002] Cat face or dog face alignment has a wide range of applications in many real-world scenarios. For example, in the process of taking pictures of pets or taking pictures with pets, by detecting the position of the pet's facial features and contour points, some texture controls or text can be added in real time. To increase the fun of shooting. However, unlike face alignment, pets are lively and active, resulting in various postures and expressions, and the variety of pets increases the difficulty of recognition. [0003] At present, for the recognition or detection method of cat face or dog face image, one is to detect the feature points of cat face directly based on the face detection model of convolutional neural networ...

AI Technical Summary

Problems solved by technology

However, unlike face alignment, pets are lively and active, resulting in various postures and expressions, and the variety of pets increases the difficulty of recognition

[0003] At present, for the recognition or detection method of cat face or dog face image, one is to detect the feature points of cat face directly based on the face detection model of convolutional neural network, and the accuracy is low; the other is to use more complex convolutional neural network Although deep learning can achieve high accuracy, it has a large amount of calculation and low efficiency; one is to use tracking or detection algorithms to detect cat faces before cat face alignment, and tracking or detection algorithms often have false detections situation, which makes the application of cat face alignment in actual shooting scenes still limited

Images