Name: Dhyey Manish Rajani
Type: User
Company: Robotics@University of Michigan Ann Arbor
Bio: Be stupid, just once, really often
Location: Ann Arbor, Michigan <---> Mumbai, India
Blog: https://in.linkedin.com/in/dhyey-manish-rajani-1629881ab
Dhyey Manish Rajani's Projects
Developed and implemented a regularized 6D pose estimation pipeline based on poseCNN architecture for generalized pose estimation in wild
Path Planning extensive implementation from scratch
A curated list of Causality in Computer Vision
Automatic colorization using deep neural networks. "Colorful Image Colorization." In ECCV, 2016.
This method uses unsupervised and soft self-supervision for auditory Martian terrain classification , based on data from the Perseverance Rover. Next Up, we plan to make this system completely self-supervised with martian terrain images
Wrote all filter-based mobile robot localization algorithms from scratch and put them under one roof i.e. here, I have (also) developed an ecosystem to bind any localization filter based python script with a customized robot motion framework in ROS.
OpenSTL: A Comprehensive Benchmark of Spatio-Temporal Predictive Learning
A Robust, light-weight and unique 3D object detection architecture providing results (better than the conventional architectures) in real-time autonomous driving scenarios
Developed a missile guidance and ballistic control system for autonomous mid-air and surface-to-air target chase and interception
Developed and implemented 2D and 3D Pose Graph SLAM using the GTSAM library and Gauss Newton Solver on the Intel and Parking Garage g2o datasets respectively
Developed and implemented all the occupancy mapping based sensor models for mobile robot mapping and successfully implemented the BKI Semantic mapping pipeline.
Regularized and extrapolative season-to-season transfer for scene adaptation
Twilight SLAM is unique framework augmenting the SLAM navigation frameworks with low-light image enhancement modules for navigating in dusky or extremely low-light or any illumination rendered featureless environments.