Study on the Deep Neural Network of Intelligent Image Detection and the Improvement of Elastic Momentum on Image Recognition

摘要

Aiming at deep neural network of intelligent image detection, this paper proposes parameter learning method of exponential elastic momentum, applies this method in the pedestrian detection, the detection accuracy is 97.8%. In addition, this algorithm is compared with adaptive momentum methods, standard momentum gradient method and elastic momentum, and the accuracy obtained through this method increases by 2.6% and 6.5% on average respectively.