Research On Logo Recognition Based On Deep Learning

With the development of computer vision technology,there will be more application of image recognition technology in real life.Meanwhile,Logo is a significant symbol of the corporate information,and logo recognition in images and videos has become a key issue,which has a wide range of applications,such as copyright infringement detection,contextual advertise placement,vehicle logo for intelligent traffic-control systems,automated computation of brand-related statistics on social media,augmented reality,etc.However the traditional algorithm in performance and efficiency is far from being able to meet these applications.With the development of artificial intelligence,deep learning has become one of the major breakthroughs in the field of artificial intelligence in recent years.Most research of traditional object detection issues can be moved in deep learning.The deep learning algorithm shows a great advantage over traditional algorithms.For the moment,the convolution neural network(CNN)method in deep learning has achieved remarkable results in image classification,face recognition and pedestrian detection,and has been widely applied.This article mainly introduces the exploration and research on Logo recognition based on deep learning.Firstly,the research status of traditional algorithm in Logo recognition and the research status of deep study convolution neural network in recent years are introduced.On this basis,we conduct the deep learning experiments.Simultaneously,we design a set of the web display system based on experimental results.The main work of this paper are following:1?We analyze the common feature model and common classifier in traditional algorithm,and make a feasibility analysis based on the issue,and introduce the neural network,and expound the advantages of CNN in image classification.2?We focus on the development of CNN algorithm over the years,and choose the faster-rcnn algorithm which with more balanced performance.For the data set we collect,we use the caffe framework which one of the deep learning platform for network training,optimize the tuning,improve the recognition accuracy,and obtain a better training model.3?According to the system requirements analysis and experimental results,we use nodejs to development web display platform.This page system is finished the function that quickly locates and recognizes the types of logo by uploading local pictures.The experimental results show that,based on the experimental platform which built on Linux,logo recognition test accuracy rate has reached 96%,and the mean average precision also reached 0.8217.In conclusion,we design and implement of web display system in the function and page performance are also in line with the needs of practical applications.