One of the main reasons of death among women worldwide is—Breast Cancer. People who have died from this life-threatening decease are more than from any other illness, including malaria or TB. A lower dose X-ray image of breast is taken to see the interior breast tissues; which is referred to mammography. It is among the best methods for identifying breast cancer. Compared to earlier equipment, mammograms exposed the breast to substantially lower radiation exposures. It has shown as a trustworthy screening instrument in recent years and an important technique for the early diagnosis of breast cancer. This work focusses on a new breast cancer detection (BCD) scheme, in which, the input image is processed using median filter (MF). Then, FCM model is deployed for segmentation. Subsequently, features like google net (deep features) and statistical features are derived that are finally subjected for the classification using Improved LinkNet (ILNet). This work focused on a new BCD scheme, in which, the input image was processed using MF. Then, FCM model was deployed for segmentation. Subsequently, features like google net (deep features) and statistical features are derived that are finally, classified using ILNet. From analysis, ILNet's accuracy was excellent. While traditional Link Net, Bi-LSTM, CNN, DBN, and SVM models achieved a less precise accuracy of 0.88, 0.89, 0.91, 0.92, and 0.84 correspondingly, ILNet employing the DDSM dataset achieved a precision of 0.98.