This repo contains the projects: 'Virtual Normal', 'DiverseDepth', and '3D Scene Shape'. They aim to solve the monocular depth estimation, 3D scene reconstruction from single image problems.

flow_tools

3D transformations for Python.

3-D Scientific Visualization in the Jupyter Notebook

OpenMMLab optical flow toolbox and benchmark

ECCV2022 - Real-Time Intermediate Flow Estimation for Video Frame Interpolation

[CVPR'22] NICE-SLAM: Neural Implicit Scalable Encoding for SLAM

Self-supervised temporally consistent depth estimation

A collection of 3D reconstruction papers in the deep learning era.

[CVPR 2022] Thin-Plate Spline Motion Model for Image Animation.

[ICLR 22] Latent Image Animator: Learning to Animate Images via Latent Space Navigation

An awesome PyTorch NeRF library

[CVPR 2020] 3D Photography using Context-aware Layered Depth Inpainting

[ACM MM 2021] Joint Implicit Image Function for Guided Depth Super-Resolution

[ECCV 2022] SimpleRecon: 3D Reconstruction Without 3D Convolutions

[ICCV 2019] Monocular depth estimation from a single image

[ECCV 2022] Map-free Visual Relocalization: Metric Pose Relative to a Single Image

Provides everything needed for high performance data loading and augmentation in pytorch.

PyTorch Lightning Optical Flow models, scripts, and pretrained weights.

Code release for "COTR: Correspondence Transformer for Matching Across Images"(ICCV 2021)

Mitsuba 3: A Retargetable Forward and Inverse Renderer

Direct voxel grid optimization for fast radiance field reconstruction.

Plenoxels: Radiance Fields without Neural Networks

Latent Point Diffusion Models for 3D Shape Generation

Official code repository for the paper: “GenSDF: Two-Stage Learning of Generalizable Signed Distance Functions”

A data generation pipeline for creating semi-realistic synthetic multi-object videos with rich annotations such as instance segmentation masks, depth maps, and optical flow.

A pipeline to reconstruct 3d meshes based on Neural Radiance Fields