id
int64 0
1
| dataset
stringclasses 1
value | scene_name
stringclasses 2
values | video_suffix
stringclasses 1
value | question
stringclasses 2
values | question_type
stringclasses 1
value |
|---|---|---|---|---|---|
0 | scannetpp | a8bf42d646 | .mp4 | [Task]
This video captures an indoor scene. Your objective is to identify specific objects within the video, understand the spatial arrangement of the scene, and estimate the center point of each object, assuming the entire scene is represented by a 10x10 grid.
[Rule]
1. We provide the cared categories: ['trash can', 'outlet', 'wall clock', 'pipe', 'fire extinguisher', 'switch', 'table', 'cable', 'helmet', 'hammer', 'notice', 'pallet', 'cart'] in this scene. Focus ONLY on these categories.
2. Estimate the center location of each instance within the provided categories. The center position should be an integer coordinate within the 10x10 grid. Ensure each object has a unique location.
3. If a category contains multiple instances, include all of them.
4. Each object's estimated location should accurately reflect its real position in the scene, preserving the relative spatial relationships among all objects.
[Output]
Present the estimated center locations for each object as a list within a dictionary.
STRICTLY follow this JSON format:
{"category name": [(x_1, y_1), ...], ...} | center_generation |
1 | scannetpp | 13c3e046d7 | .mp4 | [Task]
This video captures an indoor scene. Your objective is to identify specific objects within the video, understand the spatial arrangement of the scene, and estimate the center point of each object, assuming the entire scene is represented by a 10x10 grid.
[Rule]
1. We provide the cared categories: ['cable tray', 'wooden plank', 'pipe', 'light switch', 'jerrycan', 'rods', 'billboard', 'trash can', 'tubelight'] in this scene. Focus ONLY on these categories.
2. Estimate the center location of each instance within the provided categories. The center position should be an integer coordinate within the 10x10 grid. Ensure each object has a unique location.
3. If a category contains multiple instances, include all of them.
4. Each object's estimated location should accurately reflect its real position in the scene, preserving the relative spatial relationships among all objects.
[Output]
Present the estimated center locations for each object as a list within a dictionary.
STRICTLY follow this JSON format:
{"category name": [(x_1, y_1), ...], ...} | center_generation |