A fast and robust OSRAD filter for telemedicine applications

Abstract

Noise removal is a fundamental preprocessing step. Denoising image sequences have a significant real-time positive impact on telemedicine fields. The main objective is to improve the quality of resultant frames. This work introduces an anisotropic diffusion filter. The latter works properly when denoising images with multiplicative noise. Nevertheless, this filter cannot be appropriate for real-time implementation due to its algorithmic complexity. The purpose of our study is to achieve a high-performance gain for the anisotropic diffusion filter utilizing a graphics processor unit. It is demonstrated by the experimental results that our presented method is so effective for real-time video processing. The denoising result achieves up to 30 frames per second for a video sequence of the size of $960\times 720$ pixels. To evaluate performance, some metrics such as PSNR, SSIM, FOM, and IQI are calculated. Compared to other denoising methods, our results show a good quality of video frame, which is very interesting in the medical area.

Keywords:

• Anisotropic diffusion
• denoising
• parallel processing
• telemedicine applications
• fast implementation

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Amira Hadj Fredj

Amira Hadj Fredj a PhD student in Electronics and Microelectronics in the Faculty of Sciences of Monastir (since 2014). Her fundamental license and MS degree in Microelectronics from the Higher Institute of Informatics and Mathematics of Monastir, (Tunisia 2011 and 2014). Her main research includes SOC implementation, parallel architecture and medical image processing.

Jihene Malek

Jihene Malek a professor of electronics, she got her MS and PhD in Electronics and Microelectronics from the Faculty of Sciences (Monastir in 1996 and 2002). She has been a senior lecturer in Electronics and Microprocessors with the Electronics Department, The Higher Institute of Applied Sciences and Technology (Sousse 2015). Her current research interests include classification, classifiers combination, features extraction, biomedical engineering and medical image processing.