A new adaptive threshold method for FMDFB Gaussian subbands for image denoising application

作者：Leavline E Jebamalar^*; Sutha S; Singh D Asir Antony Gnana

来源：Optoelectronics and Advanced Materials-Rapid Communications, 2015, 9(9-10): 1315-1321.

• Image denoising
• FMDFB
• Wavelet transform
• Adaptive threshold
• SSIM

Denoising is the process of mitigating the undesired noisy components present in images. This paper proposes a new subband adaptive threshold technique in FMDFB domain for noise removal. FMDFB is a multiscale perfect reconstruction filter bank structure that is efficiently used to represent images in various image processing applications including denoising. Since, the FMDFB coefficients differ from wavelet coefficients, a mere wavelet based fixed threshold will oversmooth the images. Hence, this paper proposes a subband adaptive threshold estimation scheme to determine the suitable threshold in FMDFB domain for denoising. Our experimental results depict that the proposed denoising scheme outperforms conventional multiscale denoising schemes such as wavelet and contourlet based denoising.

• 出版日期2015-10
• 单位北京理工大学