Vivek Pawar

Developed 3D a pulmonary lung nodule detection network with deep-convolutional neural networks which achieved the sensitivity of 0.9 and 0.3 False positive per scan, tested on 491 Indian CT scans. Trained a 3D CNN to segment Lung and Lung Lobes with a 0.97 and 0.93 dice coefficient respectively. Worked On normal/abnormal chest X-ray detection and a triaging project which improved teleradiology reporting time by 35%. Managed a team of 4-6 radiologists throughout different project phases. Coordinated with data-engineer to design Azure storage structures to facilitate easier data maintenance and access across different projects.

Developed 3D a pulmonary lung nodule detection network with deep-convolutional neural networks which achieved the sensitivity of 0.9 and 0.3 False positive per scan, tested on 491 Indian CT scans. Trained a 3D CNN to segment Lung and Lung Lobes with a 0.97 and 0.93 dice coefficient respectively. Worked On normal/abnormal chest X-ray detection and a triaging project which improved teleradiology reporting time by 35%. Coded internal library for ease of training and functions regularly used in training phases.

Reproduced the code for FaceNet paper in Tensorflow to train a face recognition network. Worked on the development of Deep CNN for person tracking by creating feature embeddings. Developed POC web app for existing machine and deep learning projects.