Face retrieval method based on multitask convolution nerve network

Abstract

The invention discloses a face retrieval method based on a multitask convolution nerve network; the method comprises the following steps: detecting a random face image so as to obtain face position and key point positions; preprocessing the face image; pre-building the multitask convolution nerve network, and training the same; inputting the preprocessed face image into the trained multitask convolution nerve network; pre-building a face feature database; calculating identity feature similarity for the face image and the face feature database, thus obtaining a candidate face image list; calculating a plurality of attribute feature expression vector similarities between the face image and the candidate face images; carrying out normalization and fusing; ranking the images according to the fuse similarity scores, thus obtaining a retrieval result. The method can realize face image high quality identification, thus fast and effectively determining user IDs corresponding to mass face images, and satisfying people demands on the face identification function.

Description

Technical field [0001] The invention relates to the technical fields of artificial intelligence, pattern recognition and digital image processing, and particularly relates to a face retrieval method based on a multi-task convolutional neural network. Background technique [0002] At present, with the continuous development of human science and technology, face recognition technology is becoming more and more popular in people’s daily lives. Whether in artificial intelligence research or public safety applications, face recognition technology has always been a cutting-edge and popular technology, and it has an important role. Status. [0003] As a type of biometric recognition technology, face recognition has good development and application prospects due to its non-contact and convenient collection characteristics. Face recognition technology has played a very important role in many application scenarios, such as airport security check, border inspection and customs clearance. Wi...

AI Technical Summary

Problems solved by technology

[0005] However, with the advent of the era of big data, the scale of data that people need to process is often very large

As the capacity of the database increases, the possibility of similar faces appearing in the database will increase accordingly. When face recognition is performed based on the existing face recognition methods, the probability of misidentification will be greatly increased as the faces are verified one by one. increase, unable to meet people's fast and accurate recognition requirements for face recognition, and the overall work efficiency and quality of face recognition are low, which reduces people's experience

Images