Samarth Brahmbhatt

About

Samarth Brahmbhatt is a computer vision and robotics researcher whose work spans object detection, robotic grasping, and human-robot interaction. He is best known for his pioneering contributions to 3D object detection and pose estimation, with his 2014 paper on single-image 3D object detection and grasp-oriented pose estimation accumulating over 227 citations — a testament to its lasting influence on the field. His early work also helped democratize computer vision education through "Practical OpenCV" (2013, 57 citations), a widely adopted hands-on resource that introduced thousands of practitioners to OpenCV's capabilities. Brahmbhatt's research has progressively deepened our understanding of how robots sense and interact with the physical world. His ContactPose dataset advanced the study of hand-object contact during grasping, while ContactGrasp explored functional multi-finger grasp synthesis. More recently, he has tackled cutting-edge challenges including haptics-based manipulation, visual pressure estimation for soft robotic grippers, and diffusion-driven visual servoing. His work on collective robot learning through OpenBot-Fleet reflects a growing commitment to scalable, accessible robotics systems. Across his career, Brahmbhatt has consistently bridged perception and physical manipulation, making him a notable contributor to embodied AI and dexterous robotics research.

Research Focus

Key Achievements

Top Papers

1. 1
   Single image 3D object detection and pose estimation for grasping

   227 citations · 2014
2. 2
   Practical OpenCV

   57 citations · 2013
3. 3
   ContactPose: A Dataset of Grasps with Object Contact and Hand Pose

   14 citations · 2020
4. 4
   Introduction to Computer Vision and OpenCV

   13 citations · 2013
5. 5
   Visual Pressure Estimation and Control for Soft Robotic Grippers

   7 citations · 2022
6. 6
   Imagine2Servo: Intelligent Visual Servoing with Diffusion-Driven Goal Generation for Robotic Tasks

   5 citations · 2024
7. 7
   Identification and Learning-Based Control of an End-Effector for Targeted Throwing

   5 citations · 2024
8. 8
   Zero-Shot Transfer of Haptics-Based Object Insertion Policies

   5 citations · 2023
9. 9
   ContactGrasp: Functional Multi-finger Grasp Synthesis from Contact

   5 citations · 2019
10. 10
    OpenBot-Fleet: A System for Collective Learning with Real Robots

    1 citations · 2024

Key Collaborators

Contact & Links