Abstract participated in LSVRC-2010 contest classify 1.2 million high-resolution images into the 1000 different classes on the test data, acheived top-1 and top-5 error rate of 37.5% and 17.5% architecture has 60 million parameters and 650,000 neurons consists of 5 Conv layer, 3 FC layer Used non-saturating neuron and GPU implementation 1. Introduction 지금까지의 machine learning method를 이용한 object r..