CHANDHINII BHARATHIDASAN

“Automatic Segmentation for multi-parametric MRI”- Aiding Physicians in Accurate Tumor Detection. Preprocessed the 3D MRI kidney images. nnUNet PyTorch model - Kidney volume, cortex and medulla segmentation. 90% volume segment accuracy and 80 % cortex and medulla segment accuracy. Multi-fold ensemble model with better performance compared to single-fold models. Analyzed metrics such as DC, IoU, TP, TN, FP, FN, Sensitivity, Specificity and conducted Pearson and Bland-Altman plots with figure analysis.

Project 1 | Age and Gender detection - Motion sickness estimation Project 2 | Air leakage detection - Binary Audio Classification