Merav Keidar

MSc Student  ·  Computer Science  ·  Technion

I am advised by Or Litany (Technion) and Rajalakshmi Nandakumar (Cornell Tech). My research is in 3D computer vision, neural scene representation, and radar perception for autonomous systems.

mekeidar@gmail.com / GitHub / LinkedIn

Merav Keidar

Publications

VideoMDM

VideoMDM: Towards 3D Human Motion Generation From 2D Supervision

Amir Mann, Gal Michael Harari, Merav Keidar, Or Litany

A diffusion-based framework that trains 3D human motion priors directly from accurate 2D poses extracted from monocular videos, without any 3D ground truth.

arXiv  /  code  /  project page

Projects

UnderwaterEVER

UnderwaterEVER

Physics-informed underwater novel-view synthesis at real-time speed, with free scene dewatering at inference.

GitHub / Report / details

Python  ·  C++  ·  PyTorch  ·  CUDA / OptiX  ·  Neural Rendering

Diffusion Inpainting results

Diffusion Inpainting

Text-guided image inpainting pipeline extending Stable Diffusion 2 with four algorithmic improvements and a full evaluation framework.

GitHub / Report / details

Python  ·  PyTorch  ·  Stable Diffusion 2  ·  Generative AI

MoP Drawing Machine

MoP – Drawing Machine

A real-life pintograph drawing machine paired with a Grasshopper simulator and interactive GUI.

GitHub / details

Grasshopper  ·  Processing  ·  Generative Design  ·  Fabrication

UR5e dual-arm RRT* motion planning simulation

Robotic Motion Planning

Dual-arm UR5e manipulation using RRT* planning in configuration space, with collision avoidance, inverse kinematics, and real robot execution.

GitHub / details

Python  ·  RRT*  ·  Inverse Kinematics  ·  UR5e  ·  Robotics

Graphics GUI – reflective cow with environment mapping

Graphics GUI

Real-time 3D rendering GUI built with OpenGL, featuring configurable shaders, mesh loading, and interactive lighting.

GitHub / details

C++  ·  OpenGL  ·  ImGui  ·  GLSL