Abhishek Banerjee
About Abhishek Banerjee
I currently work as a Computer Vision Engineer at Velodyne Lidar where my experience lies in Detection and Segmentation, LiDAR-based localization pipelines, calibration pipelines, including multi-LiDAR, LiDAR-camera and LiDAR-vehicle. I also have prior experience at Nokia Bell Labs as a Deep Learning Intern where I worked with multi-rate sensor fusion for localization, multi-object tracking algorithms using deep learning-based object detector and motion modeling in crowded environments. I have worked on applications of Computer Vision and Deep Learning through strongly demonstrated projects in areas such as 3D Computer Vision, Image Segmentation, Object Detection, Image Classification and Transformer Networks. With advanced engineering experience developing computational imaging systems at the edge, I am passionate about collaborating with robotics hardware, systems, mapping and localization, and machine learning teams. Some of the areas I'm currently working on include Deep Learning inference, nonlinear optimization, registration algorithms and LiDAR-SLAM, among others. I obtained an ME in Robotics from University of Maryland, College Park MD and a BE in Mechanical Engineering from BITS Pilani, India. While at the University of Maryland, my research was in the domain of gesture/gait-based emotion detection and gait generation from text using Deep Learning methods and was done under the guidance of Professor Aniket Bera. Skills: SLAM, Visual Perception, Deep Learning, LiDAR-based localization, Languages: C/C++, Python, MATLAB, CUDA Software: Unity, Blender, SolidWorks, MS Office Suite, CUDA Framework: PyTorch, Tensorflow/Keras, Docker, ROS, OpenCV, Git Relevant Coursework: Visual Learning and Recognition , Perception for Robots, Autonomous Robotics, Planning for Robots, Computer Processing of Pictorial Information, Robot Modeling, Control of Robotic Systems Publications: • Learning Unseen Emotions from Gestures via Semantically-Conditioned Zero-Shot Perception with Adversarial Autoencoders [AAAI 2022] • Text2Gestures: A Transformer-Based Network for Generating Emotive Body Gestures for Virtual Agents [IEEE VR 2021] • Influence of flexible fins on vortex-induced load over a circular cylinder at low Reynolds number [PoF 2021]