Official PyTorch implementation of VoxFormer [CVPR 2023 Highlight]

We present MocapNET, a real-time method that estimates the 3D human pose directly in the popular Bio Vision Hierarchy (BVH) format, given estimations of the 2D body joints originating from monocular color images. Our contributions include: (a) A novel and compact 2D pose NSRM representation. (b) A human body orientation classifier and an ensemble of orientation-tuned neural networks that regress the 3D human pose by also allowing for the decomposition of the body to an upper and lower kinematic hierarchy. This permits the recovery of the human pose even in the case of significant occlusions. (c) An efficient Inverse Kinematics solver that refines the neural-network-based solution providing 3D human pose estimations that are consistent with the limb sizes of a target person (if known). All the above yield a 33% accuracy improvement on the Human 3.6 Million (H3.6M) dataset compared to the baseline method (MocapNET) while maintaining real-time performance

[CVPR 2022] "MonoScene: Monocular 3D Semantic Scene Completion": 3D Semantic Occupancy Prediction from a single image

Fast-BEV: A Fast and Strong Bird’s-Eye View Perception Baseline

Create 3d rooms in blender from floorplans.

Official project website for the CVPR 2020 paper (Oral Presentation) "Cascaded deep monocular 3D human pose estimation wth evolutionary training data"

[IROS2024] SSCBench: A Large-Scale 3D Semantic Scene Completion Benchmark for Autonomous Driving

[CVPR 2017] Generation and reconstruction of 3D shapes via modeling multi-view depth maps or silhouettes

An extension for Pixelorama that generates 3D voxel art out of 2D pixel art.

42 project cub3d, with a main menu, settings, sound & music, a minimap, actual gameplay with a random generated map, an endscreen with a highscore and more features coming soon

The official project website of "3D Human Pose Lifting with Grid Convolution" (GridConv for short, oral in AAAI 2023)

[ICRA 2024] SLCF-Net: Sequential LiDAR-Camera Fusion for Semantic Scene Completion using a 3D Recurrent U-Net

Extruding a 32-bit color image into a 3D mesh.

Unity application to convert 2D sketches to 3D models which could be maneuvered around using hand gestures (to a position and orientation of choice) in a 3D scene.

Pose Estimation for Interactive Metaverse Fitness (Capstone Design 2022)

Offboard Occupancy Refinement with Hybrid Propagation for Autonomous Driving

Using NeRF we can convert 2d images into 3d.

Create 3D side-by-side and red/cyan anaglyph videos and images from any 2D source.

This Python script reconstructs 3D models from 2D images. It uses a pre-trained deep learning model for depth estimation and Open3D for 3D processing, generating a point cloud and a 3D mesh as output.

vid23d Scallop transforms 2D video into 3D stereo SBS video using depth estimation and stereo pair generation. Utilizing deep learning and computer vision, it supports frame processing, audio merging, and enhanced visualization.

Web-based tool that turns 2D images into smooth 3D models.

Create interactive 3D Gaussian Splat scenes from a single photo — entirely in your browser.

Concatenate a sequence of 2D images (slices) e.g. jpeg into a 3D stack (volume) e.g. Nifti (ITK-based)

Tesla-like realtime road segmentation and visualization.

This repository contains code for the experimentation that I did to test the pre-trained model provided in PRNet.

AI-driven 2D-to-3D architectural rendering platform using React + TypeScript + Puter.js. Integrates multiple AI models for photorealistic visualization with serverless workers, durable storage, persistent metadata, and a real-time community feed.

Emergent Dimension

🚗 Detect vehicles in Flash Lidar datasets with advanced deep learning models, enabling robust benchmarking and innovative research in autonomous systems.

Real-Time View Synthesis with Multiplane Image Network using Multimodal Supervision

Examples and workflows for Flash LiDAR point cloud processing using MATLAB. This repository demonstrates vehicle detection in Flash LiDAR data through deep learning models for object detection and semantic segmentation in both 2D and 3D. Includes sample code for point cloud operations, training pipelines, and a public dataset available for download