In this paper, a novel method is used for the automatic detection of breast cancer by using UWB. Breast cancer is the severe threat occurs especially in women. To slacken the death rate, diagnosis and detection is a significant concern needs to be done accurately. This proposed work works well in identifying the tumor by using new algorithm and approaches. To quantize the acquired image into samples, thresholdbased segmentation is applied. The obtained input image will be preprocessed by using median filter. Median filter clear out the noise present in the UWB image. So denoising of image is needed to uphold the quality of image by noise suppression. Quality of image and feature extraction algorithm becomes unreliable due to the presence of noise. After quantization, the feature extraction is performed by using GLCM and then it is optimized. Finally Convolutional Neural Network is performed for classifying the extracted feature and it analogizes the test data with trained data. To prove its effectiveness, it is compared with other existing works, it generates a high accuracy. The accuracy achieved in this proposed work is 94%.
Key words: Ultra wide-band(UWB), Median filter, Threshold based segmentation, Gray-level co-occurrence matrix (GLCM), Convolutional Neural Network(CNN).
|
