hardik01shah commited on
Commit
125a296
·
verified ·
1 Parent(s): ba4507b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +8 -7
README.md CHANGED
@@ -63,7 +63,7 @@ wildos/
63
  | `RGB_rectified/` | Rectified (undistorted) RGB images | 390 |
64
  | `SAM_boundaries/` | Binary masks from SAM-2 boundary detection | 390 |
65
 
66
- > **Note:** Some images do not have corresponding annotations (362 out of 390 images are annotated). Images without annotations were excluded during quality control.
67
 
68
  ## Annotation Format
69
 
@@ -81,10 +81,12 @@ Each annotation file contains a list of frontier detections with the following s
81
 
82
  | Field | Description |
83
  |-------|-------------|
84
- | `label` | Frontier type (`"frontier"` or `"image_boundary_frontier"`) |
85
  | `start` | Top-left corner `[x, y]` of the bounding box |
86
  | `end` | Bottom-right corner `[x, y]` of the bounding box |
87
 
 
 
88
  ## Usage
89
 
90
  ### Loading the Dataset
@@ -162,7 +164,7 @@ visualize_frontiers(
162
 
163
  ### Metadata Mapping
164
 
165
- The `metadata.json` file in `RGB_frames/` maps each image to its source in the GrandTour dataset:
166
 
167
  ```python
168
  import json
@@ -170,8 +172,8 @@ import json
170
  with open("wildos/RGB_frames/metadata.json", "r") as f:
171
  metadata = json.load(f)
172
 
173
- # Find original GrandTour image for a specific frame
174
- original_path = metadata["rgb_00000.png"]
175
  print(f"Original GrandTour path: {original_path}")
176
  ```
177
 
@@ -179,9 +181,8 @@ print(f"Original GrandTour path: {original_path}")
179
 
180
  - **Total Images**: 390
181
  - **Total Annotations**: 362 JSON files
182
- - **Frontier Types**: `frontier`, `image_boundary_frontier`
183
  - **Image Format**: PNG
184
- - **Source Dataset**: [GrandTour Dataset](https://huggingface.co/datasets/leggedrobotics/grand_tour_dataset)
185
 
186
  ## Related Resources
187
 
 
63
  | `RGB_rectified/` | Rectified (undistorted) RGB images | 390 |
64
  | `SAM_boundaries/` | Binary masks from SAM-2 boundary detection | 390 |
65
 
66
+ > **Note:** Some images do not have corresponding annotations (362 out of 390 images are annotated). Images without annotations were excluded during quality control. The `SAM_boundaries/` folder contains SAM-2 boundary masks used in an ablation study, where frontiers were defined as the SAM boundary segments within human-annotated bounding boxes.
67
 
68
  ## Annotation Format
69
 
 
81
 
82
  | Field | Description |
83
  |-------|-------------|
84
+ | `label` | Frontier label (currently `"frontier"` for all annotations) |
85
  | `start` | Top-left corner `[x, y]` of the bounding box |
86
  | `end` | Bottom-right corner `[x, y]` of the bounding box |
87
 
88
+ > **Note:** The `label` field exists because we initially experimented with labeling frontiers of varying strengths. In the final dataset, all annotations use the single label `"frontier"`.
89
+
90
  ## Usage
91
 
92
  ### Loading the Dataset
 
164
 
165
  ### Metadata Mapping
166
 
167
+ The `metadata.json` file in `RGB_frames/` maps each image index to its source path in the GrandTour dataset:
168
 
169
  ```python
170
  import json
 
172
  with open("wildos/RGB_frames/metadata.json", "r") as f:
173
  metadata = json.load(f)
174
 
175
+ # Find original GrandTour image for a specific frame index
176
+ original_path = metadata["0"] # e.g., "release_2024-11-03-07-57-34/hdr_front/hdr_front_01342.png"
177
  print(f"Original GrandTour path: {original_path}")
178
  ```
179
 
 
181
 
182
  - **Total Images**: 390
183
  - **Total Annotations**: 362 JSON files
184
+ - **Frontier Label**: `frontier`
185
  - **Image Format**: PNG
 
186
 
187
  ## Related Resources
188