SuperGlue: Learning Feature Matching with Graph Neural Networks running through and training model process

开源项目：GitHub - HeatherJiaZG/SuperGlue-pytorch: [SuperGlue: Learning Feature Matching with Graph Neural Networks] This repo includes PyTorch code for training the SuperGlue matching network on top of SIFT keypoints and descriptors.

Paper Notes: CVPR Paper Notes｜SuperGlue bzdww

1. Download anaconda: Conda User Guide - Programmer Sought
2. Create a python3.8 environment in anaconda, and install it in this environment according to the open source project
3. Python 3
4. PyTorch >= 1.1
5. OpenCV >= 3.4 (4.1.2.30 recommended for best GUI keyboard interaction, see this note)
6. Matplotlib >= 3.1
7. NumPy >= 1.18
8. environment,

Download the dataset wget http://images.cocodataset.org/zips/train2014.zip, unzip

Configuration Environment

conda create --name py38 python=3.8

source activate py38

pip3 install numpy opencv-python torch matplotlib tqdm scipy

training model

python train.py  --train=./train2014/

问题1：ValueError: Expected more than 1 value per channel when training

Reason: Fundamentally, this is a problem when the number of features is small

Solution

try:
  desc0 = desc0 + self.kenc(kpts0, torch.transpose(data['scores0'], 0, 1))
except:
  desc0 = desc0 + self.kenc(kpts0, data['scores0'])
try:
  desc1 = desc1 + self.kenc(kpts1, torch.transpose(data['scores1'], 0, 1))
except:
  desc1 = desc1 + self.kenc(kpts1, data['scores1'])

Question 2: module 'cv2' has no atttibute 'xfeatures2d'

Reason: xfeatures2d.cpp has a patent and needs to install the corresponding version of opencv-contrib

Solution

Install the corresponding version of opencv-contrib

pip install opencv-python==4.7.0.72

pip install opencv-contrib-python==4.7.0.72