Face detection method and storage medium

Abstract

A face detection method and a storage medium, the method comprises the following steps: step (102), inputting a batch of identical training images respectively to a lightweight network and a complex network; step(104), filtering a output result of a classification figure of the lightweight network and the complex network by adopting the difficult sample mining method; step(106), constructing a comprehensive loss function, wherein the comprehensive loss function comprises a knowledge distillation loss function or a label-based face detection loss function, and the knowledge distillation loss function is obtained according to the output result of the classification figure of the lightweight network and the complex network; step(108), based on the loss function, updating parameters of the lightweight network and not updating parameters of the complex network; and step(110), repeating the above steps until the lightweight network is trained to convergence. A neural network algorithm for fast parameter adjustment is provided to solve the problem of face detection by the lightweight neural network.

Description

technical field [0001] The invention belongs to the technical field of image processing and pattern recognition, and specifically relates to a face detection method, which can be applied to security monitoring, human-computer interaction and many other fields. Background technique [0002] Face detection is an important technique, which is in demand in many computer vision applications, such as face tracking, face alignment, face recognition, etc. In recent years, due to the development of convolutional neural networks, the performance of face detection has been significantly improved. However, existing face detection models are usually computationally slow because they require relatively large neural networks to maintain good face detection performance. Although there are also one-step-based detection frameworks proposed to speed up detection (such as SSD, YOLO), they are still not fast enough for practical application scenarios, especially in CPU-based environments. On t...

AI Technical Summary

Problems solved by technology

[0004] To this end, it is necessary to provide a new type of neural network algorithm that can apply a one-step method to improve the performance of lightweight detection models and quickly adjust parameters to solve the problem of face detection with lightweight neural networks

Images