Binary Image Denoising via Incremental Neighborhood-Based Energy Optimization: A Markov Random Field Approach

• Victor José Beltrão Almajano Martinez UFMA
• Francisco Glaubos Nunes Clímaco UFMA
• Aristófanes Corrêa Silva UFMA

Resumo

A digital image may exhibit undesired noise variations, making it necessary to apply preprocessing methods to reduce noise while preserving the original structure. In this context, this paper proposed a Markov random field-inspired energy approach for optimizing binary noise reduction, considering a variable neighborhood set. Our algorithm combines Iterated Conditional Modes with Markovian prior information about the image to compute the energy associated with each pixel based on its local neighborhood, which is progressively expanded at each iteration. Experimental results on binary images corrupted with 10% random noise demonstrate that the proposed approach achieves an agreement of up to 99.85% with the original noisefree images. Future research will focus on incorporating more prior knowledge about the image to enhance the energy model, as well as its extension to grayscale images, along with the exploration of new metrics and configuration strategies.

Referências

K. Akhade, S. Ghodekar, V. Kapse, A. Raykar, and S. Wadhvane, “A survey on image denoising techniques,” International Journal of Innovative Science and Research Technology (IJISRT), vol. 9, no. 2, pp. 688–689, February 2024. [Online]. Available: [link]

M. Kopytek, P. Lech, and K. Okarma, “Application of binary image quality assessment methods to predict the quality of optical character recognition results,” Applied Sciences, vol. 14, no. 22, 2024.

R. Rodríguez, “Binarization of medical images based on the recursive application of mean shift filtering: Another algorithm,” Advances and Applications in Bioinformatics and Chemistry, vol. 1, pp. 1–12, 2008, pMCID: PMC3169934, PMID: 21918602. [Online]. Available: [link]

A. K. Jain and S. G. Nadabar, Markov Random Field Applications in Image Analysis. Boston, MA: Springer US, 1992, pp. 39–50. [Online]. DOI: 10.1007/978-1-4615-3388-7_5

M. Cherukuri, “Comparing image segmentation algorithms,” in 2024 IEEE 4th International Conference on Data Science and Computer Application (ICDSCA). IEEE, Nov. 2024, p. 266–269. [Online]. DOI: 10.1109/ICDSCA63855.2024.10859911

K. Haris, S. Efstratiadis, N. Maglaveras, and A. Katsaggelos, “Hybrid image segmentation using watersheds and fast region merging,” IEEE transactions on image processing : a publication of the IEEE Signal Processing Society, vol. 7, pp. 1684–99, 02 1998.

S. Geman and D. Geman, “Stochastic relaxation, gibbs distributions, and the bayesian restoration of images,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. PAMI-6, no. 6, 1984.

J. Besag, “On the statistical analysis of dirty pictures,” Journal of the Royal Statistical Society. Series B (Methodological), vol. 48, no. 3, pp. 259–302, 1986. [Online]. Available: [link]

R. Dubes, A. Jain, S. Nadabar, and C. Chen, “Mrf model-based algorithms for image segmentation,” in [1990] Proceedings. 10th International Conference on Pattern Recognition, 1990, pp. 808–814 vol.1.

L. Deng, “The mnist database of handwritten digit images for machine learning research [best of the web],” IEEE Signal Processing Magazine, vol. 29, no. 6, pp. 141–142, 2012.