mihaidusmanu/d2-net

Issues

• Question about the image undistortion

  #94 opened 2 years ago by ewrfcas
  7

• The link to "d2-net/hpatches-sequences-cache" seems broken.

  #108 opened a month ago by Ma0zii
  2

• unable to resolve host address ‘dsmn.ml’

  #107 opened 3 months ago by NaNtaisuike
  3

• Issue with preprocessing undistorted images; overlap matrix can not be calculated

  #100 opened a year ago by gelaretaherian
  2

• please provide source code to export onnx file

  #106 opened 7 months ago by ChengLiu-xsy
  2

• D2Net model weights

  #105 opened 6 months ago by sfm-user-1
  3

• Unable to download weight files!

  #104 opened 7 months ago by ChengLiu-xsy
  2

• unable to download d2_tf.pth

  #102 opened 9 months ago by xiaoyugit111
  3

• Undistorted and preprocessed images for our dataset

  #95 opened 2 years ago by Anuradha-Uggi
  1

• Inquiry about D2Net for Brain Tumor Segmentation on BraTS2018 Dataset

  #99 opened 9 months ago by MaggieLSY
  1

• Failed to download models

  #91 opened 2 years ago by farhadinima75
  6

• d2-net Feature extraction

  #101 opened 10 months ago by 18709817706
  0

• Error in model retraining parameters

  #98 opened a year ago by langlangmingyue
  0

• Extracting features encountered RuntimeError: indices should be either on cpu or on the same device as the indexed tensor (cpu)

  #97 opened a year ago by pfvaldez
  1

• Uploading weights to Hugging Face

  #96 opened 2 years ago by NimaBoscarino
  0

• undistorted reconstructions dataset link Error 404

  #93 opened 2 years ago by youngyyp
  2

• MegaDepth saved npz poses as 1x4 array but should be 4x4

  #92 opened 2 years ago by MACILLAS
  4

• Changing model architecture when test (inference)

  #82 opened 2 years ago by ComputerMath
  2

• Does the model_test need to train once again?

  #86 opened 2 years ago by xiezhiqing0112
  3

• Details about how to find correspondences in megadepth dataset

  #89 opened 2 years ago by gleefe1995
  0

• -

  #90 opened 2 years ago by Parskatt
  0

• How to generate the dataset for retained the network(our data include infared images and RGB images)

  #60 opened 4 years ago by zxp771
  7

• D2Net to kornia

  #84 opened 3 years ago by ducha-aiki
  3

• Improving Performance on my dataset

  #85 opened 3 years ago by ufukefe
  2

• how to transfer keypoints get from d2-net to keypoint in opencv？I want to homography and perspectiveTransform operations

  #88 opened 3 years ago by yatouxingren
  0

• How to get the image pairs for localization of Aachen dataset?

  #61 opened 4 years ago by GabbySuwichaya
  9

• missing scene_info-file in your google drive folder

  #83 opened 3 years ago by Hai-Yen
  1

• Training on custom datasets

  #81 opened 3 years ago by sachitha-bandara
  1

• why need this transformation in the extract_feature.py

  #80 opened 4 years ago by fantuslk
  1

• The custom_demo runs very slowly

  #70 opened 4 years ago by yuhf19
  3

• I encounter a problem when I run on the Fountain,Herzjesu,southbuilding dataset use d2_pth.tf model

  #77 opened 4 years ago by fantuslk
  5

• The results of Superpoint

  #76 opened 4 years ago by pixelsmaker
  1

• Loss is NaN

  #78 opened 4 years ago by csyhy1986
  2

• some correspondence points of the scene_info are outside of the images

  #75 opened 4 years ago by Hai-Yen
  4

• Distances decreasing after every consecutive input

  #73 opened 4 years ago by Aniket-Gujarathi
  9

• HPatches evaluation by SuperPoint and RootSIFT

  #74 opened 4 years ago by coper16
  1

• How to manually give the correspondences to train the network

  #72 opened 4 years ago by Aniket-Gujarathi
  2

• Are undistorted images more aligned with depth maps?

  #71 opened 4 years ago by zehongs
  4

• Training on custom dataset is detecting wrong correspondences

  #68 opened 4 years ago by UditSinghParihar
  4

• Could you help explain about the numerical results in benchmark vs the paper.

  #69 opened 4 years ago by GabbySuwichaya
  5

• Could you advice on the setting for localization on challenging conditions ?

  #62 opened 4 years ago by GabbySuwichaya
  9

• HardDetectionModule and HandcraftedLocalizationModule

  #65 opened 4 years ago by fantuslk
  1

• why the loss is always 1

  #67 opened 4 years ago by OldAAAA
  1

• Is it normal for training to take several days？

  #66 opened 4 years ago by OldAAAA
  2

• What is `bbox` field return by dataset class

  #64 opened 4 years ago by UditSinghParihar
  1

• request the experiment code

  #63 opened 4 years ago by fantuslk
  4

• loss is nan after 12 epochs

  #57 opened 4 years ago by bozhenhhu
  3

• Training Time Inquiry

  #56 opened 4 years ago by SwagJ
  4

• Some image pairs show a large number of incorrect matches

  #59 opened 4 years ago by tudipffmgt
  5

• About training another dataset

  #58 opened 4 years ago by Jo-won
  2