Using Image Retrieval to perform Computer-Aided Diagnosis of Mammographic Masses

The best method to detect breast cancer in its early stage is by using Mammographic masses. Mammography is an advanced type of x-ray used for obtaining detailed images of the breast. Several techniques may be used to solve this problem. But, many of the techniques fall short of quantifiability in the retrieval stage, so the accuracy is limited. To defeat this disservice, we tend to propose an ascendible strategy for recovery and diagnosing of mammographic bounty. In particular, for an inquiry mammographic region of interest (ROI), scale-invariant feature transform (SIFT) choices are extricated and looked in an exceedingly vocabulary tree, that stores all the quantal choices of prior analyzed mammographic ROIs. Furthermore, to absolutely apply the discriminative force of SIFT alternatives, talk data inside of the vocabulary tree is utilized to decide the weights of tree hubs. The recovered ROIs are then sent to confirm regardless of whether the inquiry ROI contains a mass. The resulting technique has an excellent quantifiability as a result of low spatial transient cost.