Image classification algorithm based on LTS-HD multi instance multi label RBF
Min Jie, Hong Zhang · 2017
In order to improve the accuracy of image classification and the robustness of the algorithm, this paper proposes a Image classification algorithm based on LTS-HD(Least Trimmed Square Hausdorff) multi instance multi label RBF. The image classification algorithm based on the traditional Hausdorff distance has poor stability, and the classification results are quite different. Therefore, the image classification algorithm based on the LTS Hausdorff overcomes the problems existing in the traditional algorithm. In this paper, an improved K-Means clustering algorithm which combines Canopy algorithm and K-Means algorithm not only optimizes the initial clustering center of K-Means algorithm, but also reduces the time complexity of traditional K-Means algorithm. The experimental results show that the proposed algorithm can effectively resist the influence of noise and improve the accuracy of classification.