A method and device for constructing a large-scale face pattern analysis sample library

Abstract

The invention discloses a method and device for constructing a large-scale human face pattern analysis sample library. The method includes: acquiring pictures in batches through the network; using the Viola-Jones face detection algorithm to perform face detection and screening on the acquired pictures; The face ratio normalizes the pictures containing the faces; the PCA+LDA algorithm is used to classify the cropped pictures by gender; the similarity comparison of the gray histogram of the pictures is used to eliminate duplicate pictures; the face samples are processed as needed The library was subjected to secondary screening and manual labeling, and then stored to local storage. The device includes a network communication interface, a display device, an input device, a memory and a processor, wherein the processor executes the method for constructing a large-scale face pattern analysis sample library. The present invention greatly reduces the labor intensity of scientific researchers building a face sample library for face pattern analysis, improves the speed of building a sample library, and the present invention can also be used for building sample libraries for other pattern recognition-related scientific research.

Description

technical field [0001] The invention relates to a method and device for constructing a large-scale human face pattern analysis sample library, and belongs to the technical field of computer image processing and pattern recognition. Background technique [0002] In the context of big data and artificial intelligence, how to acquire, store and process large-scale data is a very challenging problem. In the process of studying intelligent algorithms and engineering practice, experimental samples are an important tool for training and evaluating intelligent algorithms, and the scale of experimental samples is even more important than the exquisite design of algorithm models. At present, researchers and engineers in related fields generally build face databases manually. The process is time-consuming and labor-intensive, which severely limits the scale and expansion of the sample database. [0003] Today's sample libraries for face pattern analysis are generally small in size. O...

AI Technical Summary

Problems solved by technology

On the one hand, the problem caused by a sample library that is too small is that the algorithm cannot be fully trained and the experimental results are not objective and reliable enough.

Images