Research And Implementation Of Human Skin Information Data Analysis Method Based On Cloud Platform

With the development of Internet of things(IoT)technology and big data analysis technology,there are more and more human's skin quality detection devices on the market.People pay more and more attention to the quality of their skin.The whole process of skin quality detection will generate various kinds of skin quality information data,such as the skin and image files,the collected user information and so on.These data show an explosive growth trend.How to effectively store the data of skin quality information and build a good and extensible skin image database with good reading and writing performance is called the hot spot of social research.Skin image file of the mass,generally use the distributed storage mode,now many companies and research institutions have also put forward some skin distributed image storage solutions,but the skin image storage method of massive existing,there are some problems:(1)the existing distributed storage scheme can well support the massive small skin image file storage,waste storage resources;(2)did not consider the relationship between skin image,data storage is distributed;(3)good enough to support skin quality analysis program,affect the performance of skin quality analysis program.In order to build a distributed storage system for skin and image with good read and write performance,we store and manage skin image data efficiently,and optimize the existing distributed storage system.Based on distributed cloud storage technology and convolution neural network image classification technology,this paper designs and implements a distributed storage system for skin images.The main work of this paper in the following aspects:(1)this paper analyzes the characteristics of skin image data,propose a storage scheme for relevance classification,filing and storage of small files with skin image based on the image file to solve the skin problems of distributed storage system,improve the reading performance,and puts forward the skin image according to the classification of multi dimension cluster storage,for each copy of the project cluster between the data,improve the reliability of distributed storage system of skin image and scalability;(2)image classification algorithm research is now popular,after a comprehensive analysis of the advantages and disadvantages of various image classification algorithm is presented to analyse the relationship between the use of convolutional neural skin image network classification method and the classification of skin image.(3)when studying the skin image classification algorithm of convolutional neural network,this paper analyzes the problem of fixed learning rate when using the convolution neural network model in training,and proposes an optimization plan for this problem.Finally,according to the skin image distributed storage system designed in this paper,a multi-dimensional skin image classification model based on convolution neural network is implemented and optimized,and the effect of optimization is compared.The distributed storage system designed in this paper is deployed in the actual environment,and compared with the existing skin image data storage system,in order to verify the performance improvement of this scheme in storage and reading.The experimental results show that the proposed distributed skin image storage scheme is not only functional in integrity,but also has great improvement in reading and writing performance,and it is an excellent distributed storage scheme for skin images.