Research On Methods Of Detection And Recognition Of Multi-pose Monkey Face In Complex Background

In recent years,image detection and recognition technology is a research feature in computer vision technology.Its role is to find areas of interest from the image through the computer learning the brain's thinking method.There are many uses for the current image detection and recognition technologies,such as camera surveillance,vehicle driving systems,and virtual reality.The golden monkeys studied in this article belong to Sichuan monkeys,and they mainly survive in the surrounding areas of Sichuan and Shanxi Qinling Mountains,and they are the first-class protected animals identified by the country.Due to social progress and human development,the living environment of the golden monkey was severely damaged,making the originally scarce golden monkey in a state of extinction.Therefore,in order to protect the national level animal-the golden monkey,we study its facial features and combine current image processing techniques to conduct research on the detection and recognition of monkey faces,which helps us to better detect golden monkeys and protect them.This paper mainly focuses on monkey face detection and recognition method with pose variations under complex background.1.A method for golden monkey face detection based on regional color quantification and extended Haar-Like features is proposed in this paper.First,the original images are transformed from RGB color space to HSI color space.Second,monkey body areas are segmented from the complex background using the regional color quantization method,based on which,candidate monkey face areas are determined.Finally,combining the improved Haar-Like features with Adaboost cascade classifier,the monkey face areas are detected accurately from the candidate monkey face regions.Experimental results show that the method can not only detect frontal monkey faces well,but also improves the detection rate of monkey faces with pose variations..2.A golden monkey face recognition method based on improved CNN data enhancement is proposed.First,the data enhancement method can be used to preprocess the monkey face database,which not only increases the diversity of the image of each golden monkey,but also improves the data set of the monkey face library.Secend,combining the convolution neural network(CNN)and the twin network structure,and a new network structure that can automatically extract similar features is obtained.The network structure can be used to train the monkey face data set.Finally,the differential depth metric learning algorithm(DDML)is used as an optimization algorithm to improve the effectiveness of monkey face feature extraction.Experimental results show that this method can be used to train and test the face data of golden monkey.Compared with the traditional PCA and Adaboost algorithm,the recognition rate of the multi attitude monkey face is higher.