I. Introduction
As artificial intelligence(AI) gradually enters people's lives, more and more attention has been drawn to the concept of deep learning. Deep learning is a branch of machine learning, which imitates the working principle of the human brain's neural networks. Through constructing multi-layered neural networks and leveraging large amounts of data and computational resources, deep learning enables machines to automatically learn and extract features to solve various problems, such as autonomous driving and handwriting recognition [1], [2]. Neural network as the core of the deep learning, it connects different neurons in a layered manner basically. Through the processing the forward and backward propagation algorithms, the neural network automatically learns and adjusts the connection weights between each layers of neurons to minimize the error with respect to the truth, thereby achieving feature extraction and classification of input data [3], [4]. Deep learning has made significant breakthroughs in various fields so far, including computer vision, natural language processing, and speech recognition. Also it has been widely applied in areas such as image recognition, speech recognition, autonomous driving, and recommendation systems. CNN (Convolutional Neural Network) and Inception (GoogLeNet) are popular deep learning architectures used for computer vision tasks.