Adaptive texture and color feature based color image compression

Abstract

High compression is essential for storage and transmission of color images in multimedia communication. Efficient utilization of bandwidth and storage space is challenging task since we deal with millions of images and videos in day to day life. Lossy adaptive compression for color images is in demand as compared to grayscale images. The discrete cosine transform (DCT) is widely used in image and video coding schemes that has shown greater degree of compression as used in JPEG coder. Redundant information in an image needs to be eliminated by adopting intelligent method. In this paper, we propose efficient color and texture feature based adaptive color image compression. Color conversion from RGB to YCbCr is performed to extract color and texture features. The extracted features are used to select non-zero (significant) DCT coefficients. The storage space and bandwidth during transmission is efficiently utilized by encoding non-zero DCT coefficients and thereby preserving texture and color information in the reconstructed image. Experimentation has been carried out on different image formats successfully. The proposed technique is simple and straight forward. A good compression has been achieved with good MSE and PSNR. Experimental results for adaptive, using all coefficients and RGB color model with 20 coefficients are computed in terms of compression ratio and quality of reconstructed image are compared. The proposed adaptive method has achieved good compression ratio by retaining color and texture features.