Code release of paper "Deep Multi-View Stereo gone wild"

Overview

Deep MVS gone wild

Pytorch implementation of "Deep MVS gone wild" (Paper | website)

This repository provides the code to reproduce the experiments of the paper. It implements extensive comparison of Deep MVS architecture, training data and supervision.

If you find this repository useful for your research, please consider citing

@article{
  author    = {Darmon, Fran{\c{c}}ois  and
               Bascle, B{\'{e}}n{\'{e}}dicte  and
               Devaux, Jean{-}Cl{\'{e}}ment  and
               Monasse, Pascal  and
               Aubry, Mathieu},
  title     = {Deep Multi-View Stereo gone wild},
  year      = {2021},
  url       = {https://arxiv.org/abs/2104.15119},
}

Installation

  • Python packages: see requirements.txt

  • Fusibile:

git clone https://github.com/YoYo000/fusibile 
cd fusibile
cmake .
make .
ln -s EXE ./fusibile
  • COLMAP: see the github repository for installation details then link colmap executable with ln -s COLMAP_DIR/build/src/exe/colmap colmap

Training

You may find all the pretrained models here (120 Mo) or alternatively you can train models using the following instructions.

Data

Download the following data and extract to folder datasets

The directory structure should be as follow:

datasets
├─ blended
├─ dtu_train
├─ MegaDepth_v1
├─ undistorted_md_geometry

The data is already preprocessed for DTU and BlendedMVS. For MegaDepth, run python preprocess.py for generating the training data.

Script

The training script is train.py, launch python train.py --help for all the options. For example

  • python train.py --architecture vis_mvsnet --dataset md --supervised --logdir best_sup --world_size 4 --batch_size 4 for training the best performing setup for images in the wild.
  • python train.py --architecture mvsnet-s --dataset md --unsupervised --upsample --occ_masking --epochs 5 --lrepochs 4:10 --logdir best_unsup --world_size 3 for the best unsupervised model.

The models are saved in folder trained_models

Evaluations

We provide code for both depthmap evaluation and 3D reconstruction evaluation

Data

Download the following links and extract them to datasets

  • BlendedMVS (27.5 GB) same link as BlendedMVS training data

  • YFCC depth maps (1.1Go)

  • DTU MVS benchmark: Create directory datasets/dtu_eval and extract the following files

    In the end the folder structure should be

    datasets
    ├─ dtu_eval
        ├─ ObsMask
        ├─ images
        ├─ Points
            ├─ stl
    
  • YFCC 3D reconstruction (1.5Go)

Depthmap evaluation

python depthmap_eval.py --model MODEL --dataset DATA

  • MODEL is the name of a folder found in trained_models
  • DATA is the evaluation dataset, either yfcc or blended

3D reconstruction

See python reconstruction_pipeline.py --help for a complete list of parameters for 3D reconstruction. For running the whole evaluation for a trained model with the parameters used in the paper, run

  • scripts/eval3d_dtu.sh --model MODEL (--compute_metrics) for DTU evaluation
  • scripts/eval3d_yfcc.sh --model MODEL (--compute_metrics) for YFCC 3D evaluation

The reconstruction will be located in datasets/dtu_eval/Points or datasets/yfcc_data/Points

Acknowledgments

This repository is inspired by MVSNet_pytorch and MVSNet repositories. We also adapt the official implementations of Vis_MVSNet and CVP_MVSNet.

Copyright

Deep MVS Gone Wild All rights reseved to Thales LAS and ENPC.

This code is freely available for academic use only and Provided “as is” without any warranty.

Modification are allowed for academic research provided that the following conditions are met :
  * Redistributions of source code or any format must retain the above copyright notice and this list of conditions.
  * Neither the name of Thales LAS and ENPC nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
Owner
François Darmon
PhD student in 3D computer vision at Imagine team ENPC and Thales LAS FRANCE
François Darmon
Hand Gesture Volume Control is AIML based project which uses image processing to control the volume of your Computer.

Hand Gesture Volume Control Modules There are basically three modules Handtracking Program Handtracking Module Volume Control Program Handtracking Pro

VITTAL 1 Jan 12, 2022
Web mining module for Python, with tools for scraping, natural language processing, machine learning, network analysis and visualization.

Pattern Pattern is a web mining module for Python. It has tools for: Data Mining: web services (Google, Twitter, Wikipedia), web crawler, HTML DOM par

Computational Linguistics Research Group 8.4k Jan 03, 2023
Yolo Traffic Light Detection With Python

Yolo-Traffic-Light-Detection This project is based on detecting the Traffic light. Pretained data is used. This application entertained both real time

Ananta Raj Pant 2 Aug 08, 2022
Distributing Deep Learning Hyperparameter Tuning for 3D Medical Image Segmentation

DistMIS Distributing Deep Learning Hyperparameter Tuning for 3D Medical Image Segmentation. DistriMIS Distributing Deep Learning Hyperparameter Tuning

HiEST 2 Sep 09, 2022
Interpolation-based reduced-order models

Interpolation-reduced-order-models Interpolation-based reduced-order models High-fidelity computational fluid dynamics (CFD) solutions are time consum

Donovan Blais 1 Jan 10, 2022
Neighborhood Reconstructing Autoencoders

Neighborhood Reconstructing Autoencoders The official repository for Neighborhood Reconstructing Autoencoders (Lee, Kwon, and Park, NeurIPS 2021). T

Yonghyeon Lee 24 Dec 14, 2022
Source code and data in paper "MDFEND: Multi-domain Fake News Detection (CIKM'21)"

MDFEND: Multi-domain Fake News Detection This is an official implementation for MDFEND: Multi-domain Fake News Detection which has been accepted by CI

Rich 40 Dec 18, 2022
Pytorch library for seismic data augmentation

Pytorch library for seismic data augmentation

Artemii Novoselov 27 Nov 22, 2022
Retinal vessel segmentation based on GT-UNet

Retinal vessel segmentation based on GT-UNet Introduction This project is a retinal blood vessel segmentation code based on UNet-like Group Transforme

Kent0n 27 Dec 18, 2022
Dense Gaussian Processes for Few-Shot Segmentation

DGPNet - Dense Gaussian Processes for Few-Shot Segmentation Welcome to the public repository for DGPNet. The paper is available at arxiv: https://arxi

37 Jan 07, 2023
Controlling Hill Climb Racing with Hand Tacking

Controlling Hill Climb Racing with Hand Tacking Opened Palm for Gas Closed Palm for Brake

Rohit Ingole 3 Jan 18, 2022
Combining Reinforcement Learning and Constraint Programming for Combinatorial Optimization

Hybrid solving process for combinatorial optimization problems Combinatorial optimization has found applications in numerous fields, from aerospace to

117 Dec 13, 2022
Demystifying How Self-Supervised Features Improve Training from Noisy Labels

Demystifying How Self-Supervised Features Improve Training from Noisy Labels This code is a PyTorch implementation of the paper "[Demystifying How Sel

<a href=[email protected]"> 4 Oct 14, 2022
A smaller subset of 10 easily classified classes from Imagenet, and a little more French

Imagenette 🎶 Imagenette, gentille imagenette, Imagenette, je te plumerai. 🎶 (Imagenette theme song thanks to Samuel Finlayson) NB: Versions of Image

fast.ai 718 Jan 01, 2023
Python Classes: Medical Insurance Project using Object Oriented Programming Concepts

Medical-Insurance-Project-OOP Python Classes: Medical Insurance Project using Object Oriented Programming Concepts Classes are an incredibly useful pr

Hugo B. 0 Feb 04, 2022
Simple Text-Generator with OpenAI gpt-2 Pytorch Implementation

GPT2-Pytorch with Text-Generator Better Language Models and Their Implications Our model, called GPT-2 (a successor to GPT), was trained simply to pre

Tae-Hwan Jung 775 Jan 08, 2023
A neuroanatomy-based augmented reality experience powered by computer vision. Features 3D visuals of the Atlas Brain Map slices.

Brain Augmented Reality (AR) A neuroanatomy-based augmented reality experience powered by computer vision that features 3D visuals of the Atlas Brain

Yasmeen Brain 10 Oct 06, 2022
Catch-all collection of generative art made using processing

Generative art with Processing.py Some art I have created for fun. Dependencies Processing for Python, see how to download/use here Packages contained

2 Mar 12, 2022
TransMVSNet: Global Context-aware Multi-view Stereo Network with Transformers.

TransMVSNet This repository contains the official implementation of the paper: "TransMVSNet: Global Context-aware Multi-view Stereo Network with Trans

旷视研究院 3D 组 155 Dec 29, 2022
Cosine Annealing With Warmup

CosineAnnealingWithWarmup Formulation The learning rate is annealed using a cosine schedule over the course of learning of n_total total steps with an

zhuyun 4 Apr 18, 2022