Education Class B3
Title: Face Verification using Few-shot Deep Learning
Instructor: Amit Sethi, IIT Bombay and Abhijeet Patil, IIT Bombay
Abstract: In this tutorial, we are going to deploy the popular face recognition algorithm FaceNet on a Jetson Nano Developer Kit. Firstly, we will talk about few-shot learning and the methodology used to train FaceNet. We will go through basic building blocks such as metric learning, triplet loss and the training procedure of FaceNet. After understanding the working of FaceNet, we will jump to hardware deployment of face recognition. We will train FaceNet using pyTorch, then we will convert the trained model to half precision (FP16) using TensorRT. We will also demonstrate camera integration to Jetson Nano. After completing all the steps, we will have a working prototype of a face recognition system ready on a portable Jetson Nano device.
Bio: Amit Sethi is a Professor of Electrical Engineering at IIT Bombay, and a Visiting Instructor of Pathology at UIC. His research group works on computer vision, deep learning, and medical image analysis. His current research is focused on extracting valuable information, such as for prognosis, using deep learning on inexpensive medical modalities. He obtained his PhD in Electrical and Computer Engineering from University of Illinois at Urbana-Champaign with a focus on computer vision and machine learning, and bachelors in Electrical Engineering from IIT Delhi.
Abhijeet Patil is a PhD student at IIT Bombay, and a deep learning engineer at Griffyn Robotech – a startup working with embedded systems, robotics and deep learning in Pune India. His research interest include deep learning, computational pathology, embedded systems, object detection, tracking, and color normalization.