Breast cancer is a complex disease, and more people have been suffering from it, especially women. The Histopathological study plays a prominent role in the accurate detection and classification of breast cancer, which will be performed by pathologists. The accurate detection of breast cancer is still required to ensure this issue, even though more researchers throughout the world have proposed many techniques for accurate detection of this cancer, and some improvement is still needed for accurate results. In this study, the image processing-based framework will be used to help the pathologist get accurate results. In this framework, two techniques have been used: anomaly detection with the SVM approach (ADSVM) and resolution adaptive network (RANet). These ADSVM techniques are used for screening the patches of images that are mislabeled by using the SUFR’s extraction and LLC encoding techniques were implemented, which improves the performance of the model. In RANet, resolution and depth are used to classify the images based on their difficulty in classification. These frameworks were used with the WDBC dataset. This framework will be used to improve the prediction and accuracy of cancer.