Optimized voxel grid
Name: Yuqin Jiao | Andrew ID: sjiao2
Optimized voxel grid
Ground truth voxel grid
Optimized point cloud
Ground truth point cloud
Optimized mesh
Ground truth mesh
I used --load_feat for question 2.4, 2.5, 3.3.
Example 1
Input RGB
Predicted voxel grid
Ground truth
Example 2
Input RGB
Predicted voxel grid
Ground truth
Example 3
Input RGB
Predicted voxel grid
Ground truth
F1@0.05 score: 75.651
Example 1
Input RGB
Predicted point cloud
Ground truth
Example 2
Input RGB
Predicted point cloud
Ground truth
Example 3
Input RGB
Predicted point cloud
Ground truth
F1@0.05 score: 77.362
Example 1
Input RGB
Predicted mesh
Ground truth
Example 2
Input RGB
Predicted mesh
Ground truth
Example 3
Input RGB
Predicted mesh
Ground truth
F1@0.05 score: 73.204
Quantitative comparison (F1@0.05):
| Representation | F1@0.05 |
|---|---|
| Voxel grid | 75.651 |
| Point cloud | 77.362 |
| Mesh | 73.204 |
Intuitive explanation justifying the comparison: the mesh representation has the best visualization, but the point cloud representation shows the open part of chair better. The point cloud model has the highest F1@0.05 (77.362), followed by the voxel grid and the mesh, trained for 52, 74 and 98 iterations respectively. The mesh topology is usually better. The point cloud has no grid or fixed topology, so is more sparse.
Hyperparameter: n_points
I trained two point cloud models that differ only in n_points (100 and 1000), both for 4000 iterations, using--load_feat on the chair dataset. Both models are evaluated on the same chair set.
| n_points | F1@0.05 |
|---|---|
| 100 | 52.565 |
| 1000 | 77.418 |
F1 plots:
n_points = 100
n_points = 1000
Visual comparison (same test example):
Input RGB
Prediction (n_points = 100)
Prediction (n_points = 1000)
Analysis:
Increasing n_points from 100 to 1000 improves F1@0.05 from 52.565 to 77.418. For the samples, with 100 points the prediction is sparse and scattered, and the chair shape is not really recognizable, while with 1000 points the points are more concentrated and the prediction clearly looks like a chair. In this experiment, more points cover the chair surface more densely and give better results both quantitatively and visually.
Assumption: how does the amount of visible information in the input view affect the reconstruction of the full 3D shape?
I selected three clearly different views of the same chair and ran the model on each. All predictions are rendered from the same fixed 45° viewpoint so that they can be compared directly. I used the point model.
View 1: Informative perspective view (oblique view where the seat, back and legs are all clearly visible)
Input RGB
Prediction (fixed 45° view)
Ground truth (fixed 45° view)
View 2: Front view (some structures overlap)
Input RGB
Prediction (fixed 45° view)
Ground truth (fixed 45° view)
View 3: Blur view (heavy occlusion, only limited geometry visible)
Input RGB
Prediction (fixed 45° view)
Ground truth (fixed 45° view)
Quantitative results (same chair, F1@0.05):
| Input view | F1@0.05 |
|---|---|
| View 1: Informative perspective view | 98.683 |
| View 2: Front view | 94.849 |
| View 3: Blur view | 94.200 |
Observations:
The F1 score of the point cloud model decreases as the input view shows less of the chair: 98.683 for the informative perspective view, 94.849 for the front view, and 94.200 for the blur view. This is consistent with the idea that a view showing more of the chair gives the model more information to reconstruct the full shape. However, the differences are small, and this test uses a single chair and a single run, so it may not conclusive based on one sample.
Quantitative comparison (F1@0.05, same chair test samples):
| Training data | F1@0.05 |
|---|---|
| One class (chair) | 59.840 |
| Three classes (chair, car, plane) | 50.797 |
Example 1: model training on class class
Input RGB
Predicted voxel grid
Ground truth
F1@0.05 score: 59.840
Example 2: model training on chair car plane classes
Input RGB
Predicted voxel grid
Ground truth
Example 3: model training on chair car plane classes
Input RGB
Predicted voxel grid
Ground truth
Example 4: model training on chair car plane classes
Input RGB
Predicted voxel grid
Ground truth
F1@0.05 score: 50.797
Analysis:
Training on three classes lowers the F1@0.05 on the same chair test set from 59.840 to 50.797. Both models were trained for the same 16k iterations, so the three-class model knows chairs less, and the generated quality is not as good as the model training on chair class. As the example 1 and 3 show, the model training on chair class has a better result.