A Computer-Aided Detection (CAD) System for the Recognition of Prostate Cancer Grounded on Classification

This work describes in detail a computer-assisted diagnosis method for examining prostate MPMRI scans for signs of malignancy after segmentation. By constructing a tumour probability map, we may identify areas of interest that are statistically likely to contain malignant tumours. We've pinpointed the areas of the prostate where these suspicious structures might be hiding. In this research, we present a framework for detecting prostate cancer in segmented images and predicting the lesion grade with the help of PIRADS guidelines, both of which improve the accuracy of the diagnosis. The PZ is not the only area of the prostate where our approach can detect cancer; the TZ and CZ can also be analysed. We've created a rule-based technique for localising suspicious spots inside the prostate's several zones, using probabilistic data on cancer's presence and mapping attributes. Classifying a lesion as either “clinically significant” or “not clinically significant” is an important step. The PIRADS scoring system is an effective method for classifying tumours in a medical setting, as it may be used to estimate the likelihood that a lesion is malignant or benign. We have applied the PIRADS criteria for confirming cancer in several geographic regions. Classification tasks are performed using the K-NN classifier, the Softmax classifier, and the Random Forest classifier that we have included in our research. Each tumour is assigned a score between 1 and 5 according to the PI-RADS criteria.