Fine-grained human face image fast retrieval method based on deep learning

Abstract

The invention provides a fine-grained human face image fast retrieval method based on deep learning. The method comprises the following steps that a deep convolution neural network model is established, and a loss layer used for calculating a loss function is added to each branch layer of the deep convolution neural network model; parameters of the deep convolution neural network model are initialized; a data set is constructed, and according to a preset proportion, pictures in the data set are randomly divided into a training set, a test set and a verification set; learning parameters of the depth convolution neural network model are set; the deep convolution neural network model is trained, and the parameters of the deep convolution neural network model are updated through a stochastic gradient descent and back propagation algorithm; the trained deep convolution neural network model is tested and is specifically subjected to a coarse-grained test and a fine-grained test, human face retrieval results are obtained, and then the image retrieval efficiency and the image retrieval precision are improved.

Description

technical field [0001] The invention relates to the technical field of image retrieval, in particular to a fast retrieval method for fine-grained face images based on deep learning. Background technique [0002] With the convenience of collecting images by mobile phones and surveillance cameras and the popularity of the Internet such as Weibo and WeChat, the scale of data has grown explosively. These data greatly increase the storage and calculation burden of computing equipment. Taking face retrieval as an example, we first need to extract real-valued features such as GIST features, LBP features, and CNN features from millions of data. Secondly, use the Euclidean distance or the inner product as the similarity distance measure, and sort the pictures in the database according to the similarity with the query picture. Storing these real-valued features requires a lot of computer memory, and calculating the Euclidean distance or inner product distance between these real-valu...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the technical problems of low efficiency and poor retrieval accuracy in image retrieval in the prior art, the present invention provides a fast retrie

Images