A Large-Scale Multimodal Car Dataset with Computational Fluid Dynamics Simulations and Deep Learning Benchmarks

[3DV-2025] Official implementation of "Controllable Text-to-3D Generation via Surface-Aligned Gaussian Splatting"

R-PointHop: A Green, Accurate and Unsupervised Point Cloud Registration Method

A multi-stage graph neural network classification pipeline for protein meshes.

Colmap camera models implemented in PyTorch

Simple Python script for testing the robust estimation of the fundamental matrix between two images with RANSAC and MAGSAC++ in OpenCV, and reproducibility across 100 runs.

PyTorch code for ICPR 2020 paper "Future Urban Scene Generation Through Vehicle Synthesis"

A mathematical Python module. Allows you to play with 3D Points, Lines and Planes.

A ROS package, which gets and publishes coordinates of the middle of the lane in the robot/car frame. It uses only image, camera calibration info, robot model and tf.

A mathematical Python module. Allows you to play with 3D Position Vectors.

🛠️ Enable accurate 3D point cloud registration using ROS and Open3D, tested on Nvidia Jetson Orin with Ubuntu 20.04 and Noetic.

Collection of GNN architectures for 3D shapes and graphs using NN Template

A foundational dive into 3D vision, this project served as my hands-on introduction to processing point clouds with Open3D. I focused on core skills like plane detection, segmentation, and understanding 3D transformations from the ground up.

🌐 Explore 3D computer vision with Open3D and RealSense for point cloud processing, plane detection, and real-time object detection.