A variable block range fractal method for image compression
Ghousia Anjum Shaik, T. Bhaskara Reddy, B. Mohammed Ismail, Mansooor Alam · International Journal of Intelligent Engineering Informatics · 2024
This paper presents and implements a variable block range fractal (VBRF) method for image compression on RGB images of different categories. The proposed technique shows improvement in compression ratio (CR), peak signal noise ratio (PSNR), similarity index (SI) and compression time (CT) on applying to digital images. Mean square error (MSE), entropy and coding redundancy issues are addressed for improvements. Standard test sample images are divided into varying blocks of three categories of maximum and minimum range (Ra) block of 16 × 4, 16 × 8 and 8 × 4 for implementation. Relative fractal affine transforms are used to form iterative ranges with varying blocks reconstructing corresponding eight inverse transforms. The proposed VBRF method is applied to set of test images like Lena, satellite urban and rural, MRI, rose, bird, Zelda, pepper, God hills, etc., and improvement in compression parameters is obtained with a rate of 8% to 10%. The results obtained on CR, PSNR, SI and CT shows the effectiveness of method in improving compression rate and noise mitigation in the compressed images. Proposed method implementation parameters are compared and validated with the other popular methods of fractal compression showing a considerable improvement in performance.