Total recall understanding traffic signs using deep hierarchical convolutional neural network
German sign recognition benchmark (GTSRB) 99.33%
Belgian traffic sign classification benchmark 99.17%
1. Introduction
In this paper, we propose a noval deep convolutional neural network architecture.It is accompanied by hierarchical structure with customized skip connections in steps. The skip connections are made with dilated convolutions with different filter sizes to increase the receptive field of the feature extractors.
2. Literature review
A Feature Extracter as classificationtechnique
B Orthodox machine learning techniques
C Deep learning approaches
3. Proposed methodology
A skip connections in convolutional neural network
B Dilated convolutional operation
C Inner residual dilated skip convolution
4 Experimentation