Image Classification Based on Sparse Coding Improvement and Sparse Depth Learning Model
Download as PDF
DOI: 10.25236/scmc.2019.095
Corresponding Author
Yunpeng Chen
Abstract
For the traditional image sparse representation algorithm, it only pays attention to the feature extraction rate and ignores the multi-scale information of the image, and for the problem that the image sparse representation process is subject to noise interference and low system robustness, sparse coding improvement and deep learning are proposed. In view of the instability of the underlying pixel features of the image and the environmental impact, which is impossible to fully reflect on the semantic information and other issues, it is proposed to combine the multi-scale coefficient representation with the deep learning network through the deep network to create a learning framework and the experimental results shows that this algorithm can improve on the image recognition performance and improve the system robustness.
Keywords
Sparse Coding; Deep Learning Model; Image Classification