Image Classification And Recognition Based On The Deep Convolutional Neural Network

With the development of the information age, there were a lot of data whose features couldn't be extracted or predicted effectively in real life. One of the core function of computer vision technology is to classify and recognize, with classification and recognition as its summit mission of object detection and object positioning. Due to image data were affected by multiple factors such as illumination, environment, angle, certain object features couldn't be established by manual coding, and it is hard for high latitude data in a computer to realize real-time object detection, object localization, classification and recognition. Therefore the higher accuracy in classification could be obtained with the help of GPU of high performance and the large scale pre-training on the super database Image Net, based on the most advanced deep convolution neural network algorithm. The complete optimization training was conducted on data sets Pascal Vision Object Classes (VOC), and the real-time object detection, object localization, classification and recognition were realized by high performance GPU of NVIDIA.