Possibilistic exponential spatial fuzzy clustering based cancer segmentation in multi-parametric prostate MRI

Cancer segmentation using multi-parametric prostate magnetic resonance imaging (mp-MRI) is more opted by biomedical engineers and researchers because of its proven capability in detecting and diagnosis the sarcoma. Additionally, mp-MRI delivers added privileges in early detection of premature stages of cancer, which further helps in timely diagnosis and controlling mortality rates worldwide. In this paper, Possibilistic Exponential Spatial Fuzzy Clustering (PESFC) approach is proposed for segmenting the region of interest (ROI) i.e. cancerous region in prostate capsule. The proposed methodology is evaluated on openly available dataset (I2CVB) considering 3.0 T three different mp prostate MRI modalities i.e., T2 weighted (T2w), Dynamic Contrast Enhanced (DCE) images and Apparent Diffusion Coefficient (ADC) Maps derived from Diffusion Weighted Images (DWI). The proposed methodology is compared with other clustering techniques (Kmeans, Fuzzy C-Means clustering (FCM), Hierarchical clustering and Intuitionistic—FCM) and it is proved that proposed approach shows best performance in comparison with other methods by achieving accuracy of 89.63% at the cost of extra computational overhead. It is concluded that proposed segmentation methodology is foremost suitable for cancerous region demarcation and act as a reliable second opinion to aid radiologist in other diagnostic procedures. The proposed method is helpful to radiologist in minimizing reading time, decreasing the risk of under -/over diagnosis, refine the need of proficiency in radiology reading and facilitate large scale screening.