Performance Evaluation of an Accessory Category Recognition System Using Deep Neural Network
Yuki Sakai, Tetsuya Oda, Makoto Ikeda, Leonard Barolli · 2016
Deep Learning also called Deep Neural Network (DNN) has a deep hierarchy that connect multiple internal layers for feature detection and recognition learning. DNNs are emerging fast and will continue to grow together with feature detection methods. In previous work, we applied deep learning for vegetable object recognition and explored the Convolutional Neural Network (CNN). In this paper, we propose an enhanced accessory category recognition system which is based on CNN. From the evaluation results, we found that for recognition learning process by CNN, seventy thousand iterations were suitable. The results of learning rate was 99.77% and recognition rate was 99.75%, respectively. We observed that our system can be applied also for accessory category recognition.