Running workflow from notebooks and several other improvements

.gitignore

@@ -1,10 +1,13 @@
 # Specific
 data/
-res/
+datasets/
+*results_*/
+assets/output/
 output/
 sandbox/
-assets/outs
 src/deep_image_matching/thirdparty/weights
+fra_*
+sb_*
 
 # VScode
 .vscode/

.vscode/launch.json

@@ -15,21 +15,25 @@
       "program": "${file}",
       "console": "integratedTerminal",
       "args": [
+        "--dir",
+        "datasets/easy_small",
+        // "datasets/belvedere_stereo",
+        // "datasets/casalbagliano",
+        // "/home/francesco/casalbagliano/C",
         "--config",
-        // "superpoint+lightglue",
-        "roma",
-        "--images",
-        "data/belv_lingua_easy",
+        "superpoint+lightglue",
+        "--strategy",
+        // "matching_lowres",
+        "bruteforce",
         "--quality",
-        "low",
+        "medium",
         "--tiling",
-        "preselection",
-        "--strategy",
-        "low",
-        // "--upright",
-        "--skip_reconstruction",
-        "--force",
+        "none",
         "-V"
+        // "--db_path",
+        // "/home/francesco/casalbagliano/B_extended/results_superpoint+lightglue_bruteforce_quality_high/reconstruction",
+        // "--skip_reconstruction",
+        // "--force",
       ]
     },
     {
@@ -39,6 +43,19 @@
       "program": "${file}",
       "console": "integratedTerminal",
       "args": ["--gui"]
+    },
+    {
+      "name": "Python: dense",
+      "type": "python",
+      "request": "launch",
+      "program": "${file}",
+      "console": "integratedTerminal",
+      "args": [
+        "--sfm_dir",
+        "/home/francesco/casalbagliano/subset_A/deep_image_matching/results_superpoint+lightglue_bruteforce_quality_high",
+        "--dense_dir",
+        "/home/francesco/casalbagliano/subset_A/deep_image_matching/results_roma_sequential_quality_low"
+      ]
     }
   ]
 }

README.md

@@ -7,8 +7,8 @@
 | ------------------------------------------------ | -------------------------------------------------- | ---------------------------------------------------- | ---------------------------------------------- |
 | <img src='assets/matches_sift.gif' height="100"> | <img src='assets/matches_joined.gif' height="100"> | <img src='assets/orientation_deep.gif' height="100"> | <img src='assets/roma_dense.gif' height="100"> |
 
-| SIFT                                             | SUPERGLUE                                            |            
-| ------------------------------------             | ------------------------------------                 |
+| SIFT                                             | SUPERGLUE                                            |
+| ------------------------------------------------ | ---------------------------------------------------- |
 | <img src='assets/temple_rsift.gif' height="165"> | <img src='assets/temple_superglue.gif' height="165"> |
 
 Multivew matcher for SfM software. Support both deep-learning based and hand-crafted local features and matchers and export keypoints and matches directly in a COLMAP database or to Agisoft Metashape by importing the reconstruction in Bundler format. It supports both CLI and GUI. Feel free to collaborate!
@@ -37,14 +37,15 @@ Key features:
 | &check; SIFT (opencv)              |
 | &check; DeDoDe                     |
 
-| Supported SfM software                       |
-| ---------------------------------------------|
-| &check; COLMAP                               |
-| &check; OpenMVG                              |
-| &check; Agisoft Metashape                    |
-| &check; Software that supports bundler format|
+| Supported SfM software                        |
+| --------------------------------------------- |
+| &check; COLMAP                                |
+| &check; OpenMVG                               |
+| &check; Agisoft Metashape                     |
+| &check; Software that supports bundler format |
+
+## Colab demo ➡️ [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/3DOM-FBK/deep-image-matching/blob/master/notebooks/deep_image_matching_example.ipynb)
 
-## Colab demo  ➡️ [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/3DOM-FBK/deep-image-matching/blob/master/notebooks/deep_image_matching_example.ipynb)
 Want to run on a sample dataset? Try the Colab demo!
 
 ## Installation
@@ -101,7 +102,7 @@ There are two main configuration in `config.py`: `conf_general` and `confs`.
       "geom_verification": GeometricVerification.PYDEGENSAC, -> `Enable or disable Geometric Verification. Options are: NONE: disabled; PYDEGENSAC: use pydegensac; MAGSAC: use OpenCV geometric verification with MAGSAC.`
       "gv_threshold": 4, -> `Threshold [px] for the geometric verification`
       "gv_confidence": 0.9999, -> `Confidence value for the geometric verification`
-      "preselection_size_max": 2000, -> `if tile_selection == TileSelection.PRESELECTION, define the resolution at which the images are downsampled to run the low-resolution tile preselection.`
+      "tile_preselection_size": 2000, -> `if tile_selection == TileSelection.PRESELECTION, define the resolution at which the images are downsampled to run the low-resolution tile preselection.`
     }
   ```
 
@@ -206,6 +207,7 @@ The GUI loads the available configurations from `config.py` and it shows them in
 ### Merging databases with different local features
 
 To run the matching with different local features and/or matchers and marging together the results, you can use scripts in the `./scripts` directory for merging the COLMAP databases.
+
 ```bash
 python ./join_databases.py --help
 python ./join_databases.py --input assets/to_be_joined --output assets/to_be_joined

config.yaml

@@ -0,0 +1,14 @@
+# Configuration file for Deep Image Matching
+
+# General
+general:
+  data_dir: /home/username/data
+  output_dir: /home/username/output
+  config: "superpoint+lightglue"
+
+extractor:
+
+
+
+matcher:
+  name: flann