TOTAL VARIATION WITH GRADIENT HISTOGRAM: A NOVEL APPROACH TO IMAGE DENOISING

Renuka Devi S M

Abstract: Images play important role in digital world for various research oriented applications and in day to day life, but these images are often available in defective conditions due to noise the gets mixed up with them while capturing and transmitting. Removal of noise from these defective images for better quality plays crucial role, and can be done by using various methods. Natural images have high total variation in them, that is, the integral of the absolute gradient of the signal is high. Image denoising done using Non-linear Total variation (TV) method smoothens the denoised image. To overcome this, a novel gradient histogram method is included in TV method to preserve the fine details of image. Experiments are conducted with various texture structures of image and proved that this method improves the quality of image when compared to earlier method.

Keywords: Total Variation, Total Variation Gradient Histogram, Signal Noise Ratio, Structural SIMilarity (SSIM), Feature Similarity (FSIM)

DOI: https://doi.org/10.15623/ijret.2016.0511002