Application of SPCA Algorithm in Image Dimensionality Reduction
Xianwei Wu, WenYang Yu, Yu-Bin Yang · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2014
In the page, We discuss several dimensionality reduction methods for image feature, and then focus on the one: SPCA(Simple Primary Component Analysis), which is simple fast and exceeding algorithm of data-oriented PCA algorithm.In order to better understand the SPCA algorithm, Some well-designed experiments of image compression and image retrieval are taken to compare these algorithms.By experiment 1, we get the result: PCA matrix algorithm is best in performance but worst in speed, and GHA is better in speed ,but worst in performance, and the results show that SPCA is out-standing not only in performance, but also in speed.By experiment 2, we get the desired result: using the image feature after SPCA almost get the same performance of original image feature, but much better than original image feature in speed.The conclusion is: SPCA algorithm can be applied in many field, especially in image compression and image retrieval.