Enhancing Mammography Images with Artificial Intelligence to Improve Radiological Diagnosis in Breast Cancer

Öz

Breast cancer is one of the most common types of cancer in women, and early diagnosis is life-saving. The aim of this study is to enhance the resolution of mammography images, thereby improving the contrast resolution, spatial resolution, and the detectability of calcifications, distortions, and opacities in the images. For this purpose, mammography images obtained from the open-access mini-MIAS dataset were used. Both the original dataset and the images processed with the CLAHE (Contrast Limited Adaptive Histogram Equalization) algorithm underwent resolution enhancement using the Stable Diffusion artificial intelligence system. The results were evaluated by an expert radiologist, and it was determined that the diagnostic quality of the images significantly increased. These improvements aim to support early diagnosis in breast cancer and enhance diagnostic accuracy. Additionally, the applicability and effectiveness of these methods were emphasized, and the potential benefits of resolution enhancement techniques in clinical practice were discussed. The results have the potential to allow for more detailed and accurate analysis of mammography images, thereby improving patient care and treatment planning.

Anahtar Kelimeler

• Mammography
• Image processing
• Resolution enhancement
• Artificial intelligence
• Clahe algorithm

Kaynakça

1. Al-Najdawi, Nijad, Mariam Biltawi, Sara Tedmori. 2015. Mammogram image visual enhancement, mass segmentation and classification. Applied Soft Comput, 35: 175–85. doi:10.1016/J.ASOC.2015.06.029
2. Avcı, Hanife, Jale Karakaya. 2023. A novel medical image enhancement algorithm for breast cancer detection on mammography images using machine learning. Diagnostics, 13(3): 348. doi:10.3390/DIAGNOSTICS13030348
3. Besl, Paul J, Ramesh C Jain. 1988. Segmentation through variable-order surface fitting. IEEE Transact Pattern Analys Mach Intellig, 10(2): 167–92. doi:10.1109/34.3881
4. Dhawan, Atam P, Yateen S. Chitre, Myron Moskowitz, Eric Gruenstein. 1991. Classification of mammographic microcalcification and structural features using an artificial neural network. https://researchwith.njit.edu/en/publications/classification-of-mammographic-microcalcification-and-structural- (accessed date: August 12, 2024).
5. Ganvir, Neha N, DM Yadav. 2019. Filtering method for pre-processing mammogram images for breast cancer detection. Inter J Engin Adv Technol, 9(1): 4222–29. doi:10.35940/ijeat.A1623.109119
6. Guide. 2024. What denoising strength does and how to use it in stable diffusion. https://onceuponanalgorithm.org/guide-what-denoising-strength-does-and-how-to-use-it-in-stable-diffusion/ (accessed date: August 12, 2024).
7. Li, Hongjun, Shanhua Zhang, Qingyuan Wang, Rongguang Zhu. 2016. Clinical value of mammography in diagnosis and identification of breast mass. Pakistan J Med Sci, 32(4): 1020. doi:10.12669/PJMS.324.9384.
8. Mehdy MM, PY Ng, EF Shair, NI Md Saleh, C Gomes. 2017. Artificial neural networks in image processing for early detection of breast cancer. Comput Math Meth Med, 2017(1): 2610628. doi:10.1155/2017/2610628

9. Ramani, R, N Suthanthira Vanitha, S Valarmathy. 2013. The pre-processing techniques for breast cancer detection in mammography images. Inter J Image, Graph Signal Proc, (5): 47–54. doi:10.5815/ijigsp.2013.05.06
10. Rombach, Robin, Andreas Blattmann, Dominik Lorenz, Patrick Esser, Bjorn Ommer. 2022. High-resolution image synthesis with latent diffusion models. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 5-6 June, Dubai, UAE, pp: 10674–106785. doi:10.1109/CVPR52688.2022.01042
11. Steins. 2023. Stable Diffusion clearly explained! Medium. http://anakin.ai/blog/stable-diffusion-sampling-steps/ (accessed date: August 12, 2024).
12. Swathi C, Anoop BK. Anto Sahaya Dhas D, Perumal Sanker S. 2017. Comparison of different image preprocessing methods used for retinal fundus images. 2017 Conference on Emerging Devices and Smart Systems, ICEDSS 2017: 175–79. doi:10.1109/ICEDSS.2017.8073677