文章基本信息

标题：Performance Comparison of Image Classification using KFCG with Assorted Pixel Window Sizes in RGB and LUV Color Spaces
本地全文：下载
作者：Dr. H. B. Kekre ; Dr. Sudeep D. Thepade ; Varun K. Banura 等
期刊名称：International Journal of Computer Science & Technology
印刷版ISSN：2229-4333
电子版ISSN：0976-8491
出版年度：2012
卷号：3
期号：2
页码：164-168
语种：English
出版社：Ayushmaan Technologies
摘要：The theme of the work presented in the paper is Vector Quantization (VQ) based image classification using Kekre’s Fast Codebook Generation (KFCG) algorithm for variable size pixel windows in RGB and LUV color spaces. Here different codebooks of size 8, 16, 32, 64, 128, 256 and 512 generated using KFCG for pixel windows of size 1×2, 1×3, 2×1, 2×2, 3×1 and 3×3 are considered as feature vectors and are used in image classification. These feature vectors are calculated for RGB and LUV color spaces. The proposed techniques of content based image classification are tested on a general image database of 1000 images consisting of 11 different categories. The proposed image classification methods give higher success rate with pixel windows of smaller size indicating that increase in size of pixel window leads to over fitting of the image classifier. Also when the proposed KFCG based image retrieval techniques were compared in RGB and LUV color spaces, it was observed that the performance in LUV color space was ameliorated. Among all the proposed techniques, codebook of size 32 with pixel window size 1×3 in LUV color space gives the best performance.