A Lightweight Residual-Inception Convolutional Neural Network
Kangyu Gao, Qingyong Zhang, Haoran Wang · Journal of Physics Conference Series · 2019
Abstract Deep convolutional neural networks have become a powerful tool to solve practical problems, especially in the field of image recognition and machine learning. This paper introduces a lightweight convolutional neural network model which named RINet, based on a combination of the Inception structure and blocks inspired by ResNet. This model achieves high accuracy while reducing the number of parameters and increases the training speed. It has reached the correct rate of 93.7% on the dataset of ISIC and the recognition accuracy improves 6.3% and 1.5% compared to deeper networks called InceptionV1 and transfer learning of InceptionV3.