Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
  • Loading branch information
simone-fontana authored Dec 4, 2023
1 parent 7b43c09 commit 797b457
Showing 1 changed file with 0 additions and 168 deletions.
168 changes: 0 additions & 168 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,171 +7,3 @@ This work compares different neural network-based algorithms for point clouds re
The results of the registration problems can be found in the ```results``` folder, divided by algorithm used. In each leaf folder, a file with the ```_config.json``` suffix can be found, stating the parameters used for the registration.

The full instructions to reproduce the experiments can be found in the [wiki](https://github.com/iralabdisco/neural_registration_comparison/wiki) of this repository.

## Results
These are the results of different approaches. Feature extractors have been used in conjunction with RANSAC, FastGlobal, and TEASER. Results are reported as the median, 0.75 quantile, and 0.95 quantile of the residual error (as defined in the [Point Clouds Registraion Benchmark](https://github.com/iralabdisco/point_clouds_registration_benchmark).

### 3DFeatNet features
| | RANSAC | | | FastGlobal | | | TEASER | | |
|----------------------------------|--------|-------:|-------:|------------|-------:|-------:|--------|-------:|-------:|
| | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q |
|** Sequence** | | | | | | | | | |
|**plain** | 8.40 | 12.24 | 18.73 | 4.99 | 6.17 | 7.61 | 9.57 | 13.26 | 22.57 |
|**stairs** | 2.17 | 3.09 | 7.07 | 2.31 | 3.36 | 5.26 | 2.30 | 3.78 | 7.39 |
|**apartment** | 1.92 | 2.78 | 4.81 | 2.03 | 2.53 | 3.73 | 2.22 | 3.27 | 5.10 |
|**hauptgebaude** | 2.72 | 6.47 | 9.28 | 2.79 | 4.91 | 6.31 | 2.44 | 5.57 | 8.72 |
|**wood\_autumn** | 3.20 | 4.42 | 5.92 | 2.53 | 3.62 | 5.10 | 2.87 | 4.10 | 5.95 |
|**wood\_summer** | 2.79 | 3.73 | 5.34 | 2.21 | 3.52 | 4.17 | 2.49 | 3.83 | 5.06 |
|**gazebo\_summer** | 3.51 | 4.81 | 8.41 | 2.33 | 3.30 | 4.17 | 3.13 | 4.98 | 7.15 |
|**gazebo\_winter** | 3.72 | 5.47 | 9.10 | 2.15 | 3.34 | 4.51 | 3.14 | 4.92 | 8.09 |
|**box\_met** | 10.14 | 12.76 | 19.82 | 8.30 | 10.53 | 13.98 | 13.01 | 28.65 | 81.78 |
|**p2at\_met** | 10.43 | 17.48 | 32.84 | 6.27 | 10.27 | 15.89 | 10.22 | 18.40 | 89.83 |
|**planetary\_map** | 41.00 | 60.25 | 100.93 | 27.08 | 42.20 | 61.59 | 40.76 | 62.61 | 103.20 |
|**pioneer\_slam** | 10.25 | 17.67 | 35.64 | 6.74 | 8.91 | 10.64 | 13.87 | 27.94 | 43.60 |
|**pioneer\_slam3** | 4.73 | 9.13 | 13.84 | 5.86 | 7.98 | 9.47 | 7.47 | 12.44 | 21.51 |
|**long\_office\_household** | 14.93 | 27.53 | 41.75 | 4.13 | 7.77 | 11.24 | 9.68 | 28.31 | 72.94 |
|**urban05** | 2.27 | 3.54 | 6.79 | 1.33 | 2.03 | 3.47 | 2.17 | 3.71 | 7.09 |
|**Total** | 4.21 | 9.96 | 37.31 | 3.53 | 6.53 | 21.15 | 4.17 | 10.55 | 50.44 |


### FCGF features using the model trained on 3dMatch

| | RANSAC | | | FastGlobal | | | TEASER | | |
|----------------------------------|--------|-------:|-------:|------------|-------:|-------:|--------|-------:|-------:|
| | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q |
|**Sequence** | | | | | | | | | |
|**plain** | 7.05 | 8.90 | 13.55 | 7.22 | 9.19 | 13.95 | 7.22 | 9.34 | 13.82 |
|**stairs** | 5.77 | 7.45 | 9.15 | 5.85 | 7.43 | 9.28 | 5.85 | 7.50 | 9.28 |
|**apartment** | 5.93 | 7.38 | 10.94 | 6.15 | 7.62 | 10.33 | 6.15 | 7.63 | 10.33 |
|**hauptgebaude** | 2.53 | 3.08 | 3.71 | 2.59 | 3.22 | 3.78 | 2.62 | 3.18 | 3.78 |
|**wood\_autumn** | 3.19 | 4.21 | 5.64 | 3.92 | 5.12 | 5.79 | 3.84 | 5.12 | 5.79 |
|**wood\_summer** | 2.96 | 3.73 | 4.98 | 2.60 | 3.44 | 4.55 | 2.54 | 3.51 | 4.63 |
|**gazebo\_summer** | 2.84 | 4.07 | 5.06 | 3.11 | 4.35 | 5.20 | 3.14 | 4.36 | 5.44 |
|**gazebo\_winter** | 3.50 | 4.79 | 5.85 | 3.60 | 4.77 | 5.85 | 3.61 | 4.78 | 5.68 |
|**box\_met** | 7.61 | 10.84 | 13.79 | 8.40 | 12.07 | 17.89 | 8.92 | 11.43 | 17.93 |
|**p2at\_met** | 9.82 | 12.75 | 17.48 | 8.43 | 13.13 | 21.25 | 8.69 | 12.97 | 19.59 |
|**planetary\_map** | 11.33 | 15.52 | 21.16 | 31.09 | 46.49 | 67.82 | 31.92 | 47.25 | 67.94 |
|**pioneer\_slam** | 11.74 | 18.76 | 30.26 | 11.77 | 18.76 | 30.29 | 11.77 | 18.76 | 30.30 |
|**pioneer\_slam3** | 7.07 | 12.86 | 16.62 | 7.07 | 12.86 | 16.64 | 7.06 | 12.85 | 16.63 |
|**long\_office\_household** | 9.21 | 20.70 | 41.05 | 9.41 | 20.66 | 41.05 | 9.68 | 20.65 | 41.05 |
|**urban05** | 2.51 | 3.76 | 5.30 | 1.41 | 1.93 | 3.16 | 2.27 | 3.68 | 11.04 |
|**Total** | 4.85 | 8.73 | 18.11 | 4.92 | 9.53 | 30.27 | 5.08 | 9.71 | 30.35 |

### FCGF features using the model trained on Kitti

| | RANSAC | | | FastGlobal | | | TEASER | | |
|----------------------------------|--------|-------:|-------:|------------|-------:|-------:|--------|-------:|-------:|
| | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q |
|**Sequence** | | | | | | | | | |
|**plain** | 7.05 | 8.90 | 13.55 | 7.22 | 9.19 | 13.95 | 7.22 | 9.34 | 13.82 |
|**stairs** | 5.77 | 7.45 | 9.15 | 5.85 | 7.43 | 9.28 | 5.85 | 7.50 | 9.28 |
|**apartment** | 5.93 | 7.38 | 10.94 | 6.15 | 7.62 | 10.33 | 6.15 | 7.63 | 10.33 |
|**hauptgebaude** | 2.53 | 3.08 | 3.71 | 2.59 | 3.22 | 3.78 | 2.62 | 3.18 | 3.78 |
|**wood\_autumn** | 3.19 | 4.21 | 5.64 | 3.92 | 5.12 | 5.79 | 3.84 | 5.12 | 5.79 |
|**wood\_summer** | 2.96 | 3.73 | 4.98 | 2.60 | 3.44 | 4.55 | 2.54 | 3.51 | 4.63 |
|**gazebo\_summer** | 2.84 | 4.07 | 5.06 | 3.11 | 4.35 | 5.20 | 3.14 | 4.36 | 5.44 |
|**gazebo\_winter** | 3.50 | 4.79 | 5.85 | 3.60 | 4.77 | 5.85 | 3.61 | 4.78 | 5.68 |
|**box\_met** | 7.61 | 10.84 | 13.79 | 8.40 | 12.07 | 17.89 | 8.92 | 11.43 | 17.93 |
|**p2at\_met** | 9.82 | 12.75 | 17.48 | 8.43 | 13.13 | 21.25 | 8.69 | 12.97 | 19.59 |
|**planetary\_map** | 11.33 | 15.52 | 21.16 | 31.09 | 46.49 | 67.82 | 31.92 | 47.25 | 67.94 |
|**pioneer\_slam** | 11.74 | 18.76 | 30.26 | 11.77 | 18.76 | 30.29 | 11.77 | 18.76 | 30.30 |
|**pioneer\_slam3** | 7.07 | 12.86 | 16.62 | 7.07 | 12.86 | 16.64 | 7.06 | 12.85 | 16.63 |
|**long\_office\_household** | 9.21 | 20.70 | 41.05 | 9.41 | 20.66 | 41.05 | 9.68 | 20.65 | 41.05 |
|**urban05** | 2.51 | 3.76 | 5.30 | 1.41 | 1.93 | 3.16 | 2.27 | 3.68 | 11.04 |
|**Total** | 4.85 | 8.73 | 18.11 | 4.92 | 9.53 | 30.27 | 5.08 | 9.71 | 30.35 |

### FPFH

| | RANSAC | | | FastGlobal | | | TEASER | | |
|----------------------------------|--------|-------:|-------:|------------|-------:|-------:|--------|-------:|-------:|
| | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q |
|**Sequence** | | | | | | | | | |
|**plain** | 0.41 | 3.74 | 6.87 | 0.49 | 1.11 | 1.89 | 0.12 | 0.31 | 1.83 |
|**stairs** | 0.21 | 2.37 | 4.45 | 0.28 | 1.95 | 3.89 | 0.08 | 0.58 | 3.57 |
|**apartment** | 0.21 | 1.33 | 3.57 | 0.21 | 1.15 | 2.67 | 0.13 | 1.46 | 3.47 |
|**hauptgebaude** | 0.19 | 1.61 | 3.34 | 0.36 | 1.05 | 1.86 | 0.02 | 0.97 | 2.04 |
|**wood\_autumn** | 2.90 | 3.75 | 5.43 | 0.32 | 0.51 | 0.89 | 0.05 | 0.08 | 0.24 |
|**wood\_summer** | 2.39 | 3.69 | 5.23 | 0.30 | 0.47 | 1.11 | 0.04 | 0.07 | 0.17 |
|**gazebo\_summer** | 1.90 | 2.77 | 4.38 | 0.20 | 0.43 | 1.21 | 0.04 | 0.06 | 0.27 |
|**gazebo\_winter** | 1.52 | 2.97 | 5.60 | 0.18 | 0.37 | 1.25 | 0.04 | 0.08 | 0.40 |
|**box\_met** | 7.44 | 10.46 | 13.23 | 7.16 | 9.87 | 14.12 | 0.31 | 7.62 | 12.10 |
|**p2at\_met** | 8.04 | 12.35 | 17.48 | 1.90 | 6.01 | 10.05 | 0.16 | 0.66 | 9.92 |
|**planetary\_map** | 10.80 | 15.11 | 20.78 | 29.82 | 43.52 | 69.25 | 46.45 | 62.77 | 95.28 |
|**pioneer\_slam** | 0.36 | 1.57 | 12.30 | 3.24 | 7.46 | 10.79 | 0.46 | 1.92 | 7.24 |
|**pioneer\_slam3** | 0.13 | 0.19 | 0.43 | 0.13 | 0.23 | 0.64 | 0.14 | 0.21 | 0.90 |
|**long\_office\_household** | 0.48 | 1.69 | 18.07 | 1.47 | 6.91 | 12.22 | 0.66 | 1.67 | 15.04 |
|**urban05** | 2.93 | 3.89 | 5.53 | 1.53 | 2.13 | 3.44 | 2.77 | 4.07 | 6.40 |
|**Total** | 1.82 | 4.48 | 13.12 | 0.48 | 2.44 | 19.90 | 0.15 | 1.56 | 33.49 |


### 3DSmoothNet

| | RANSAC | | | FastGlobal | | | TEASER | | |
|----------------------------------|--------|-------:|-------:|------------|-------:|-------:|--------|-------:|-------:|
| | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q |
|**Sequence** | | | | | | | | | |
|**plain** | 0.04 | 0.09 | 10.33 | 1.40 | 2.22 | 3.18 | 0.03 | 1.30 | 2.83 |
|**stairs** | 0.01 | 0.03 | 5.16 | 1.08 | 2.55 | 4.95 | 0.01 | 0.03 | 4.54 |
|**apartment** | 0.01 | 0.01 | 0.06 | 0.59 | 1.67 | 3.63 | 0.01 | 0.02 | 0.04 |
|**hauptgebaude** | 0.79 | 0.85 | 2.35 | 0.76 | 1.34 | 2.00 | 0.80 | 1.58 | 2.37 |
|**wood\_autumn** | 0.02 | 0.02 | 0.05 | 0.65 | 1.46 | 1.82 | 0.01 | 0.02 | 0.02 |
|**wood\_summer** | 0.01 | 0.02 | 0.03 | 0.47 | 0.95 | 1.66 | 0.01 | 0.01 | 0.02 |
|**gazebo\_summer** | 0.01 | 0.02 | 0.08 | 0.57 | 1.47 | 3.94 | 0.01 | 0.01 | 0.02 |
|**gazebo\_winter** | 0.01 | 0.01 | 0.03 | 0.39 | 0.78 | 2.26 | 0.01 | 0.01 | 0.02 |
|**box\_met** | 7.44 | 10.75 | 13.79 | 8.69 | 10.49 | 13.89 | 7.05 | 9.90 | 13.24 |
|**p2at\_met** | 4.93 | 10.61 | 15.92 | 6.37 | 8.00 | 11.89 | 0.16 | 6.54 | 11.84 |
|**planetary\_map** | 11.08 | 15.22 | 20.95 | 29.89 | 41.09 | 66.74 | 41.08 | 64.64 | 101.85 |
|**pioneer\_slam** | 0.07 | 0.10 | 0.72 | 0.08 | 0.16 | 0.80 | 0.03 | 0.07 | 0.12 |
|**pioneer\_slam3** | 0.03 | 0.05 | 0.09 | 0.26 | 0.66 | 3.14 | 0.01 | 0.02 | 0.03 |
|**long\_office\_household** | 0.08 | 0.14 | 0.27 | 0.06 | 0.13 | 0.35 | 0.04 | 0.08 | 0.20 |
|**urban05** | 2.51 | 3.62 | 5.29 | 1.66 | 2.07 | 3.10 | 1.32 | 1.93 | 3.37 |
|**Total** | 0.04 | 1.83 | 12.28 | 0.89 | 2.60 | 18.92 | 0.02 | 0.80 | 27.31 |


### FMR using the model trained on 7scenes

| | RANSAC | | | FastGlobal | | | TEASER | | |
|----------------------------------|--------|-------:|-------:|------------|-------:|-------:|--------|-------:|-------:|
| | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q |
|**Sequence** | | | | | | | | | |
|**plain** | 0.04 | 0.09 | 10.33 | 1.40 | 2.22 | 3.18 | 0.03 | 1.30 | 2.83 |
|**stairs** | 0.01 | 0.03 | 5.16 | 1.08 | 2.55 | 4.95 | 0.01 | 0.03 | 4.54 |
|**apartment** | 0.01 | 0.01 | 0.06 | 0.59 | 1.67 | 3.63 | 0.01 | 0.02 | 0.04 |
|**hauptgebaude** | 0.79 | 0.85 | 2.35 | 0.76 | 1.34 | 2.00 | 0.80 | 1.58 | 2.37 |
|**wood\_autumn** | 0.02 | 0.02 | 0.05 | 0.65 | 1.46 | 1.82 | 0.01 | 0.02 | 0.02 |
|**wood\_summer** | 0.01 | 0.02 | 0.03 | 0.47 | 0.95 | 1.66 | 0.01 | 0.01 | 0.02 |
|**gazebo\_summer** | 0.01 | 0.02 | 0.08 | 0.57 | 1.47 | 3.94 | 0.01 | 0.01 | 0.02 |
|**gazebo\_winter** | 0.01 | 0.01 | 0.03 | 0.39 | 0.78 | 2.26 | 0.01 | 0.01 | 0.02 |
|**box\_met** | 7.44 | 10.75 | 13.79 | 8.69 | 10.49 | 13.89 | 7.05 | 9.90 | 13.24 |
|**p2at\_met** | 4.93 | 10.61 | 15.92 | 6.37 | 8.00 | 11.89 | 0.16 | 6.54 | 11.84 |
|**planetary\_map** | 11.08 | 15.22 | 20.95 | 29.89 | 41.09 | 66.74 | 41.08 | 64.64 | 101.85 |
|**pioneer\_slam** | 0.07 | 0.10 | 0.72 | 0.08 | 0.16 | 0.80 | 0.03 | 0.07 | 0.12 |
|**pioneer\_slam3** | 0.03 | 0.05 | 0.09 | 0.26 | 0.66 | 3.14 | 0.01 | 0.02 | 0.03 |
|**long\_office\_household** | 0.08 | 0.14 | 0.27 | 0.06 | 0.13 | 0.35 | 0.04 | 0.08 | 0.20 |
|**urban05** | 2.51 | 3.62 | 5.29 | 1.66 | 2.07 | 3.10 | 1.32 | 1.93 | 3.37 |
|**Total** | 0.04 | 1.83 | 12.28 | 0.89 | 2.60 | 18.92 | 0.02 | 0.80 | 27.31 |


### FMR using the model trained on Modelnet40

| | RANSAC | | | FastGlobal | | | TEASER | | |
|----------------------------------|--------|-------:|-------:|------------|-------:|-------:|--------|-------:|-------:|
| | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q | Median | 0.75 Q | 0.95 Q |
|**Sequence** | | | | | | | | | |
|**plain** | 0.04 | 0.09 | 10.33 | 1.40 | 2.22 | 3.18 | 0.03 | 1.30 | 2.83 |
|**stairs** | 0.01 | 0.03 | 5.16 | 1.08 | 2.55 | 4.95 | 0.01 | 0.03 | 4.54 |
|**apartment** | 0.01 | 0.01 | 0.06 | 0.59 | 1.67 | 3.63 | 0.01 | 0.02 | 0.04 |
|**hauptgebaude** | 0.79 | 0.85 | 2.35 | 0.76 | 1.34 | 2.00 | 0.80 | 1.58 | 2.37 |
|**wood\_autumn** | 0.02 | 0.02 | 0.05 | 0.65 | 1.46 | 1.82 | 0.01 | 0.02 | 0.02 |
|**wood\_summer** | 0.01 | 0.02 | 0.03 | 0.47 | 0.95 | 1.66 | 0.01 | 0.01 | 0.02 |
|**gazebo\_summer** | 0.01 | 0.02 | 0.08 | 0.57 | 1.47 | 3.94 | 0.01 | 0.01 | 0.02 |
|**gazebo\_winter** | 0.01 | 0.01 | 0.03 | 0.39 | 0.78 | 2.26 | 0.01 | 0.01 | 0.02 |
|**box\_met** | 7.44 | 10.75 | 13.79 | 8.69 | 10.49 | 13.89 | 7.05 | 9.90 | 13.24 |
|**p2at\_met** | 4.93 | 10.61 | 15.92 | 6.37 | 8.00 | 11.89 | 0.16 | 6.54 | 11.84 |
|**planetary\_map** | 11.08 | 15.22 | 20.95 | 29.89 | 41.09 | 66.74 | 41.08 | 64.64 | 101.85 |
|**pioneer\_slam** | 0.07 | 0.10 | 0.72 | 0.08 | 0.16 | 0.80 | 0.03 | 0.07 | 0.12 |
|**pioneer\_slam3** | 0.03 | 0.05 | 0.09 | 0.26 | 0.66 | 3.14 | 0.01 | 0.02 | 0.03 |
|**long\_office\_household** | 0.08 | 0.14 | 0.27 | 0.06 | 0.13 | 0.35 | 0.04 | 0.08 | 0.20 |
|**urban05** | 2.51 | 3.62 | 5.29 | 1.66 | 2.07 | 3.10 | 1.32 | 1.93 | 3.37 |
|**Total** | 0.04 | 1.83 | 12.28 | 0.89 | 2.60 | 18.92 | 0.02 | 0.80 | 27.31 |

0 comments on commit 797b457

Please sign in to comment.