Abstract

Breast cancer remains a leading cause of mortality among women globally. Early and accurate diagnosis using medical imaging, such as mammograms or ultrasound, is critical for effective treatment. However, challenges such as low contrast, noise, and poor image quality in raw medical datasets often hinder accurate detection and diagnosis. Many conventional image preprocessing techniques fail to enhance pathological features effectively, which are essential for early stage breast cancer recognition. Noise artifacts and blurred edges further degrade the performance of diagnostic models. This paper proposes an integrated approach that combines advanced image filtering and enhancement techniques including Gaussian Filtering, Contrast Limited Adaptive Histogram Equalization (CLAHE), and Wavelet-Based Sharpening. These are applied in sequence to reduce noise, enhance tumor boundaries, and improve Thus contrast in mammographic images. The processed images are then used to train deep learning classifiers (e.g., CNNs) to improve detection accuracy. Experimental evaluations on public breast cancer imaging datasets demonstrate a significant improvement in diagnostic accuracy, sensitivity, and precision. The enhanced images yield clearer visualization of microcalcifications and tumor regions, leading to over 93% accuracy in detection.

Authors

A. Muthumari¹, Subhash A. Nalawade²
Anna University Regional Campus Madurai, India¹, Dr. D. Y. Patil Institute of Technology, India²

Keywords

Breast Cancer, Image Enhancement, Medical Imaging, Noise Reduction, Deep Learning