ROBUST IMAGE CLASSIFICATION BASED ON FIDELITY WEIGHTED DISCRIMINANT CO-COMPETITIVE REPRESENTATION
In order to deeply mine the information between categories and improve the robustness and accuracy of the method,a robust image classification method based on weighted discriminant cooperative competition representation is proposed.The discriminant and competition-collaboration representation between all classes was integrated in a unified model.Two discriminant constraints and the constraint of weighted class representation coefficient were introduced to further enhance the contribution rate of class to representation.Furthermore,a robust algorithm with fidelity was introduced to improve the robustness to noise.Experimental results on six sets of image data show that the proposed method has higher classification accuracy and robustness.
国家科技期刊平台
NSTL
万方数据
