Improved Breast Cancer Detection using Wavelet Based Image Enhancement

Mammogram and ultrasound being the most effective imaging modality for breast cancer detection . The quality of image is to be enhanced for efficient detection of cancer. Mammogram being a low contrast image and ultrasound being indifferentiable requires image enhancement before being processed for detection of cancer. Thus, an image enhancement workflow is essential for mammogram and ultrasound images for improving the detection of breast cancer. Therefore, this work enhances breast images using bottom hat feature extraction, CLAHE, wavelet transformation and sharpening of image. The detection framework uses Deep learning networks for detection of cancer from enhanced image. The proposed methodology results in improved metrics in enhanced breast mammogram and ultrasound scan image . Further, the enhanced image shows an increased accuracy in detection of breast cancer as benign or malignant using Convolutional Neural Network (CNN) architecture.