Hugging Face's logo

Datasets:
acane2
/
GIO

ArXiv:
Tags:
video-understanding
License:
Dataset card Data Studio Files Community
1
GIO
File size: 2,493 Bytes 
8f580d3
 
 
 
 
 
 
 
 
 
f102654
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
 
---
license: mit
tags:
- video-understanding
---

Dataset of the paper [Interacted Object Grounding in Spatio-Temporal Human-Object Interactions](https://huggingface.co/papers/2412.19542)

Code: https://github.com/DirtyHarryLYL/HAKE-AVA

# Preparing Dataset for HAKE-GIO + HAKE-AVA-PaSta (h box + b box & class + action + PaSta)

1. Dataset downloading steps

    1. Download AVA Dataset (following [SlowFast](https://github.com/facebookresearch/SlowFast)). 

        ```
        ./script/download_AVA_dataset.sh
        ```

    2. Downloading annotation

        The annotation is contained in GIO_annotation

        Please download it to ava folder and extract data from the package.

    3. Structure of downloaded data

        ```
        GIO
        |_ GIO_annotation
        |  |_ GIO_test.csv
        |  |_ GIO_train.csv
        |_ frames
        |  |_ [video name 0]
        |  |  |_ [video name 0]_000001.jpg
        |  |  |_ [video name 0]_000002.jpg
        |  |  |_ ...
        |  |_ [video name 1]
        |     |_ [video name 1]_000001.jpg
        |     |_ [video name 1]_000002.jpg
        |     |_ ...
        |_ frame_lists
        |  |_ train.csv
        |  |_ val.csv
        ```

2. Annotation Format

    Files in the GIO folder contains the annotations of each frame, including human/object box, action, object name, etc.

    example:

    | video       | frame | h_x1  | h_y1  | h_x2  | h_y2  | o_x1  | o_y1  | o_x2  | o_y2  | action | object_name | human_id | object_id |
    | ----------- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ------ | ----------- | -------- | --------- |
    | -5KQ66BBWC4 | 905   | 0.392 | 0.033 | 0.556 | 0.618 | 0.37  | 0.019 | 0.432 | 0.608 | 6      | stick       | 12       | 0         |
    | -5KQ66BBWC4 | 906   | 0.408 | 0.008 | 0.586 | 0.639 | 0.37  | 0.036 | 0.457 | 0.678 | 6      | stick       | 12       | 0         |
    | -5KQ66BBWC4 | 907   | 0.42  | 0.115 | 0.616 | 0.883 | 0.371 | 0.143 | 0.466 | 0.878 | 6      | stick       | 12       | 0         |

    The meanings of each column:

    - video: name of the video
    - frame: time (second) of the frame
    - h_x1~h_y2: the upper left and bottom right corners of human-box
    - o_x1~o_y2: the upper left and bottom right corners of object-box
    - action: the action label of the person in the human-box
    - object_name: name of object
    - human_id: ID of the person performing the action
    - object_id: category id of the object