Implementasi resmi Grounding Image Matching in 3D with MASt3R
[Halaman proyek], [MASt3R arxiv], [DUSt3R arxiv]
@misc { mast3r_arxiv24 ,
title = { Grounding Image Matching in 3D with MASt3R } ,
author = { Vincent Leroy and Yohann Cabon and Jerome Revaud } ,
year = { 2024 } ,
eprint = { 2406.09756 } ,
archivePrefix = { arXiv } ,
primaryClass = { cs.CV }
}
@inproceedings { dust3r_cvpr24 ,
title = { DUSt3R: Geometric 3D Vision Made Easy } ,
author = { Shuzhe Wang and Vincent Leroy and Yohann Cabon and Boris Chidlovskii and Jerome Revaud } ,
booktitle = { CVPR } ,
year = { 2024 }
}Kode ini didistribusikan di bawah Lisensi CC BY-NC-SA 4.0. Lihat LISENSI untuk informasi lebih lanjut.
# Copyright (C) 2024-present Naver Corporation. All rights reserved.
# Licensed under CC BY-NC-SA 4.0 (non-commercial use only). git clone --recursive https://github.com/naver/mast3r
cd mast3r
# if you have already cloned mast3r:
# git submodule update --init --recursiveconda create -n mast3r python=3.11 cmake=3.14.0
conda activate mast3r
conda install pytorch torchvision pytorch-cuda=12.1 -c pytorch -c nvidia # use the correct version of cuda for your system
pip install -r requirements.txt
pip install -r dust3r/requirements.txt
# Optional: you can also install additional packages to:
# - add support for HEIC images
# - add required packages for visloc.py
pip install -r dust3r/requirements_optional.txt # DUST3R relies on RoPE positional embeddings for which you can compile some cuda kernels for faster runtime.
cd dust3r/croco/models/curope/
python setup.py build_ext --inplace
cd ../../../../Anda dapat memperoleh pos pemeriksaan dengan dua cara:
Anda dapat menggunakan integrasi huggingface_hub kami: model akan diunduh secara otomatis.
Jika tidak, Kami menyediakan beberapa model terlatih:
| Nama model | Resolusi pelatihan | Kepala | Pembuat enkode | Dekoder |
|---|---|---|---|---|
MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric | 512x384, 512x336, 512x288, 512x256, 512x160 | CatMLP+DPT | ViT-L | ViT-B |
Anda dapat memeriksa hyperparameter yang kami gunakan untuk melatih model ini di bagian: Hyperparameter kami Pastikan untuk memeriksa lisensi kumpulan data yang kami gunakan.
Untuk mengunduh model tertentu, misalnya MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric.pth :
mkdir -p checkpoints/
wget https://download.europe.naverlabs.com/ComputerVision/MASt3R/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric.pth -P checkpoints/Untuk pos pemeriksaan ini, pastikan untuk menyetujui lisensi semua kumpulan data pelatihan yang kami gunakan, selain CC-BY-NC-SA 4.0. Lisensi kumpulan data mapfree khususnya sangat terbatas. Untuk informasi lebih lanjut, periksa CHECKPOINTS_NOTICE.
Kami membuat satu ruang berpelukan yang menjalankan penyelarasan global baru yang jarang dalam demo sederhana untuk adegan kecil: naver/MASt3R Ada dua demo yang tersedia untuk dijalankan secara lokal:
demo.py is the updated demo for MASt3R. It uses our new sparse global alignment method that allows you to reconstruct larger scenes
python3 demo.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric
# Use --weights to load a checkpoint from a local file, eg --weights checkpoints/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric.pth
# Use --local_network to make it accessible on the local network, or --server_name to specify the url manually
# Use --server_port to change the port, by default it will search for an available port starting at 7860
# Use --device to use a different device, by default it's "cuda"
demo_dust3r_ga.py is the same demo as in dust3r (+ compatibility for MASt3R models)
see https://github.com/naver/dust3r?tab=readme-ov-file#interactive-demo for details
Untuk menjalankan MASt3R menggunakan Docker, termasuk dengan dukungan NVIDIA CUDA, ikuti petunjuk berikut:
Instal Docker : Jika belum terinstal, unduh dan instal docker dan docker compose dari situs web Docker.
Instal NVIDIA Docker Toolkit : Untuk dukungan GPU, instal toolkit NVIDIA Docker dari situs web Nvidia.
Bangun image Docker dan jalankan : cd ke direktori ./docker dan jalankan perintah berikut:
cd docker
bash run.sh --with-cuda --model_name= " MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric "Atau jika Anda ingin menjalankan demo tanpa dukungan CUDA, jalankan perintah berikut:
cd docker
bash run.sh --model_name= " MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric " Secara default, demo.py diluncurkan dengan opsi --local_network .
Kunjungi http://localhost:7860/ untuk mengakses UI web (atau ganti localhost dengan nama mesin untuk mengaksesnya dari jaringan).
run.sh akan meluncurkan docker-compose menggunakan file konfigurasi docker-compose-cuda.yml atau docker-compose-cpu.ym, lalu memulai demo menggunakan entrypoint.sh.
from mast3r . model import AsymmetricMASt3R
from mast3r . fast_nn import fast_reciprocal_NNs
import mast3r . utils . path_to_dust3r
from dust3r . inference import inference
from dust3r . utils . image import load_images
if __name__ == '__main__' :
device = 'cuda'
schedule = 'cosine'
lr = 0.01
niter = 300
model_name = "naver/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric"
# you can put the path to a local checkpoint in model_name if needed
model = AsymmetricMASt3R . from_pretrained ( model_name ). to ( device )
images = load_images ([ 'dust3r/croco/assets/Chateau1.png' , 'dust3r/croco/assets/Chateau2.png' ], size = 512 )
output = inference ([ tuple ( images )], model , device , batch_size = 1 , verbose = False )
# at this stage, you have the raw dust3r predictions
view1 , pred1 = output [ 'view1' ], output [ 'pred1' ]
view2 , pred2 = output [ 'view2' ], output [ 'pred2' ]
desc1 , desc2 = pred1 [ 'desc' ]. squeeze ( 0 ). detach (), pred2 [ 'desc' ]. squeeze ( 0 ). detach ()
# find 2D-2D matches between the two images
matches_im0 , matches_im1 = fast_reciprocal_NNs ( desc1 , desc2 , subsample_or_initxy1 = 8 ,
device = device , dist = 'dot' , block_size = 2 ** 13 )
# ignore small border around the edge
H0 , W0 = view1 [ 'true_shape' ][ 0 ]
valid_matches_im0 = ( matches_im0 [:, 0 ] >= 3 ) & ( matches_im0 [:, 0 ] < int ( W0 ) - 3 ) & (
matches_im0 [:, 1 ] >= 3 ) & ( matches_im0 [:, 1 ] < int ( H0 ) - 3 )
H1 , W1 = view2 [ 'true_shape' ][ 0 ]
valid_matches_im1 = ( matches_im1 [:, 0 ] >= 3 ) & ( matches_im1 [:, 0 ] < int ( W1 ) - 3 ) & (
matches_im1 [:, 1 ] >= 3 ) & ( matches_im1 [:, 1 ] < int ( H1 ) - 3 )
valid_matches = valid_matches_im0 & valid_matches_im1
matches_im0 , matches_im1 = matches_im0 [ valid_matches ], matches_im1 [ valid_matches ]
# visualize a few matches
import numpy as np
import torch
import torchvision . transforms . functional
from matplotlib import pyplot as pl
n_viz = 20
num_matches = matches_im0 . shape [ 0 ]
match_idx_to_viz = np . round ( np . linspace ( 0 , num_matches - 1 , n_viz )). astype ( int )
viz_matches_im0 , viz_matches_im1 = matches_im0 [ match_idx_to_viz ], matches_im1 [ match_idx_to_viz ]
image_mean = torch . as_tensor ([ 0.5 , 0.5 , 0.5 ], device = 'cpu' ). reshape ( 1 , 3 , 1 , 1 )
image_std = torch . as_tensor ([ 0.5 , 0.5 , 0.5 ], device = 'cpu' ). reshape ( 1 , 3 , 1 , 1 )
viz_imgs = []
for i , view in enumerate ([ view1 , view2 ]):
rgb_tensor = view [ 'img' ] * image_std + image_mean
viz_imgs . append ( rgb_tensor . squeeze ( 0 ). permute ( 1 , 2 , 0 ). cpu (). numpy ())
H0 , W0 , H1 , W1 = * viz_imgs [ 0 ]. shape [: 2 ], * viz_imgs [ 1 ]. shape [: 2 ]
img0 = np . pad ( viz_imgs [ 0 ], (( 0 , max ( H1 - H0 , 0 )), ( 0 , 0 ), ( 0 , 0 )), 'constant' , constant_values = 0 )
img1 = np . pad ( viz_imgs [ 1 ], (( 0 , max ( H0 - H1 , 0 )), ( 0 , 0 ), ( 0 , 0 )), 'constant' , constant_values = 0 )
img = np . concatenate (( img0 , img1 ), axis = 1 )
pl . figure ()
pl . imshow ( img )
cmap = pl . get_cmap ( 'jet' )
for i in range ( n_viz ):
( x0 , y0 ), ( x1 , y1 ) = viz_matches_im0 [ i ]. T , viz_matches_im1 [ i ]. T
pl . plot ([ x0 , x1 + W0 ], [ y0 , y1 ], '-+' , color = cmap ( i / ( n_viz - 1 )), scalex = False , scaley = False )
pl . show ( block = True )
Di bagian ini, kami menyajikan demonstrasi singkat untuk memulai pelatihan MASt3R.
Lihat bagian Kumpulan Data di DUSt3R
Seperti pada demo pelatihan DUSt3R, kami akan mengunduh dan menyiapkan subset CO3Dv2 yang sama - Creative Commons Attribution-NonCommercial 4.0 International dan meluncurkan kode pelatihan di dalamnya. Ini adalah proses yang sama persis seperti DUSt3R. Model demo akan dilatih selama beberapa periode pada kumpulan data yang sangat kecil. Itu tidak akan terlalu bagus.
# download and prepare the co3d subset
mkdir -p data/co3d_subset
cd data/co3d_subset
git clone https://github.com/facebookresearch/co3d
cd co3d
python3 ./co3d/download_dataset.py --download_folder ../ --single_sequence_subset
rm ../ * .zip
cd ../../..
python3 datasets_preprocess/preprocess_co3d.py --co3d_dir data/co3d_subset --output_dir data/co3d_subset_processed --single_sequence_subset
# download the pretrained dust3r checkpoint
mkdir -p checkpoints/
wget https://download.europe.naverlabs.com/ComputerVision/DUSt3R/DUSt3R_ViTLarge_BaseDecoder_512_dpt.pth -P checkpoints/
# for this example we'll do fewer epochs, for the actual hyperparameters we used in the paper, see the next section: "Our Hyperparameters"
torchrun --nproc_per_node=4 train.py
--train_dataset " 1000 @ Co3d(split='train', ROOT='data/co3d_subset_processed', aug_crop='auto', aug_monocular=0.005, aug_rot90='diff', mask_bg='rand', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], n_corres=8192, nneg=0.5, transform=ColorJitter) "
--test_dataset " 100 @ Co3d(split='test', ROOT='data/co3d_subset_processed', resolution=(512,384), n_corres=1024, seed=777) "
--model " AsymmetricMASt3R(pos_embed='RoPE100', patch_embed_cls='ManyAR_PatchEmbed', img_size=(512, 512), head_type='catmlp+dpt', output_mode='pts3d+desc24', depth_mode=('exp', -inf, inf), conf_mode=('exp', 1, inf), enc_embed_dim=1024, enc_depth=24, enc_num_heads=16, dec_embed_dim=768, dec_depth=12, dec_num_heads=12, two_confs=True) "
--train_criterion " ConfLoss(Regr3D(L21, norm_mode='?avg_dis'), alpha=0.2) + 0.075*ConfMatchingLoss(MatchingLoss(InfoNCE(mode='proper', temperature=0.05), negatives_padding=0, blocksize=8192), alpha=10.0, confmode='mean') "
--test_criterion " Regr3D_ScaleShiftInv(L21, norm_mode='?avg_dis', gt_scale=True, sky_loss_value=0) + -1.*MatchingLoss(APLoss(nq='torch', fp=torch.float16), negatives_padding=12288) "
--pretrained " checkpoints/DUSt3R_ViTLarge_BaseDecoder_512_dpt.pth "
--lr 0.0001 --min_lr 1e-06 --warmup_epochs 1 --epochs 10 --batch_size 4 --accum_iter 4
--save_freq 1 --keep_freq 5 --eval_freq 1 --disable_cudnn_benchmark
--output_dir " checkpoints/mast3r_demo "
Kami tidak merilis semua set data pelatihan, namun berikut adalah perintah yang kami gunakan untuk melatih model kami:
# MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric - train mast3r with metric regression and matching loss
# we used cosxl to generate variations of DL3DV: "foggy", "night", "rainy", "snow", "sunny" but we were not convinced by it.
torchrun --nproc_per_node=8 train.py
--train_dataset "57_000 @ Habitat512(1_000_000, split='train', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 68_400 @ BlendedMVS(split='train', mask_sky=True, resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 68_400 @ MegaDepth(split='train', mask_sky=True, resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 45_600 @ ARKitScenes(split='train', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 22_800 @ Co3d(split='train', mask_bg='rand', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 22_800 @ StaticThings3D(mask_bg='rand', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 45_600 @ ScanNetpp(split='train', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 45_600 @ TartanAir(pairs_subset='', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 4_560 @ UnrealStereo4K(resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 1_140 @ VirtualKitti(optical_center_is_centered=True, resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 22_800 @ WildRgbd(split='train', mask_bg='rand', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 145_920 @ NianticMapFree(split='train', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 57_000 @ DL3DV(split='nlight', resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 57_000 @ DL3DV(split='not-nlight', cosxl_augmentations=None, resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5) + 34_200 @ InternalUnreleasedDataset(resolution=[(512, 384), (512, 336), (512, 288), (512, 256), (512, 160)], aug_crop='auto', aug_monocular=0.005, transform=ColorJitter, n_corres=8192, nneg=0.5)"
--test_dataset " Habitat512(1_000, split='val', resolution=(512,384), seed=777, n_corres=1024) + 1_000 @ BlendedMVS(split='val', resolution=(512,384), mask_sky=True, seed=777, n_corres=1024) + 1_000 @ ARKitScenes(split='test', resolution=(512,384), seed=777, n_corres=1024) + 1_000 @ MegaDepth(split='val', mask_sky=True, resolution=(512,336), seed=777, n_corres=1024) + 1_000 @ Co3d(split='test', resolution=(512,384), mask_bg='rand', seed=777, n_corres=1024) "
--model " AsymmetricMASt3R(pos_embed='RoPE100', patch_embed_cls='ManyAR_PatchEmbed', img_size=(512, 512), head_type='catmlp+dpt', output_mode='pts3d+desc24', depth_mode=('exp', -inf, inf), conf_mode=('exp', 1, inf), enc_embed_dim=1024, enc_depth=24, enc_num_heads=16, dec_embed_dim=768, dec_depth=12, dec_num_heads=12, two_confs=True, desc_conf_mode=('exp', 0, inf)) "
--train_criterion " ConfLoss(Regr3D(L21, norm_mode='?avg_dis'), alpha=0.2, loss_in_log=False) + 0.075*ConfMatchingLoss(MatchingLoss(InfoNCE(mode='proper', temperature=0.05), negatives_padding=0, blocksize=8192), alpha=10.0, confmode='mean') "
--test_criterion " Regr3D(L21, norm_mode='?avg_dis', gt_scale=True, sky_loss_value=0) + -1.*MatchingLoss(APLoss(nq='torch', fp=torch.float16), negatives_padding=12288) "
--pretrained " checkpoints/DUSt3R_ViTLarge_BaseDecoder_512_dpt.pth "
--lr 0.0001 --min_lr 1e-06 --warmup_epochs 8 --epochs 50 --batch_size 4 --accum_iter 2
--save_freq 1 --keep_freq 5 --eval_freq 1 --print_freq=10 --disable_cudnn_benchmark
--output_dir " checkpoints/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric "
Lihat bagian Visloc di DUSt3R
Dengan visloc.py Anda dapat menjalankan eksperimen pelokalan visual kami di Aachen-Day-Night, InLoc, Cambridge Landmarks, dan 7 Scenes.
# Aachen-Day-Night-v1.1:
# scene in 'day' 'night'
# scene can also be 'all'
python3 visloc.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric --dataset " VislocAachenDayNight('/path/to/prepared/Aachen-Day-Night-v1.1/', subscene=' ${scene} ', pairsfile='fire_top50', topk=20) " --pixel_tol 5 --pnp_mode poselib --reprojection_error_diag_ratio 0.008 --output_dir /path/to/output/Aachen-Day-Night-v1.1/ ${scene} /loc
# or with coarse to fine:
python3 visloc.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric --dataset " VislocAachenDayNight('/path/to/prepared/Aachen-Day-Night-v1.1/', subscene=' ${scene} ', pairsfile='fire_top50', topk=20) " --pixel_tol 5 --pnp_mode poselib --reprojection_error_diag_ratio 0.008 --output_dir /path/to/output/Aachen-Day-Night-v1.1/ ${scene} /loc --coarse_to_fine --max_batch_size 48 --c2f_crop_with_homography
# InLoc
python3 visloc.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric --dataset " VislocInLoc('/path/to/prepared/InLoc/', pairsfile='pairs-query-netvlad40-temporal', topk=20) " --pixel_tol 5 --pnp_mode poselib --reprojection_error_diag_ratio 0.008 --output_dir /path/to/output/InLoc/loc
# or with coarse to fine:
python3 visloc.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric --dataset " VislocInLoc('/path/to/prepared/InLoc/', pairsfile='pairs-query-netvlad40-temporal', topk=20) " --pixel_tol 5 --pnp_mode poselib --reprojection_error_diag_ratio 0.008 --output_dir /path/to/output/InLoc/loc --coarse_to_fine --max_image_size 1200 --max_batch_size 48 --c2f_crop_with_homography
# 7-scenes:
# scene in 'chess' 'fire' 'heads' 'office' 'pumpkin' 'redkitchen' 'stairs'
python3 visloc.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric --dataset " VislocSevenScenes('/path/to/prepared/7-scenes/', subscene=' ${scene} ', pairsfile='APGeM-LM18_top20', topk=1) " --pixel_tol 5 --pnp_mode poselib --reprojection_error_diag_ratio 0.008 --output_dir /path/to/output/7-scenes/ ${scene} /loc
# Cambridge Landmarks:
# scene in 'ShopFacade' 'GreatCourt' 'KingsCollege' 'OldHospital' 'StMarysChurch'
python3 visloc.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric --dataset " VislocCambridgeLandmarks('/path/to/prepared/Cambridge_Landmarks/', subscene=' ${scene} ', pairsfile='APGeM-LM18_top50', topk=20) " --pixel_tol 5 --pnp_mode poselib --reprojection_error_diag_ratio 0.008 --output_dir /path/to/output/Cambridge_Landmarks/ ${scene} /loc
