Data_card.md

Data Card for WhereIsWally Object Detection

Dataset Description

Dataset Summary

The WhereIsWally dataset contains snapshots of different scenarios of the 17 existing Where's Wally books, a British series of children's puzzle books created by English illustrator Martin Handford. Each scenario contains one or more of the following characters: Odlaw, Wally, Wilma and Wizard. Per snapshot, zero, one or more of these characters can be found with corresponding annotations. This naturally facilitates building an object detection model to detect these characters for an unseen Where's Wally-scenario.

fe. A snapshot containing two characters ('Odlaw' and 'Wilma'):

Supported Tasks and Leaderboards

• Character Detection: This dataset support building an Where's Wally character(s) detection model.

Dataset Structure

Data Instances

A datapoint comprises an image (snapshots of scenario) that contains none, one or more annoted characters. An example of a (raw) data entry example with annotations (under the "boxes" field):

{
    "camera": "Generated by Roboflow",
    "datasets": [
        "niAVe8eMYrNWkWN1y3FJ"
    ],
    "destination": "00e68521cf184455e4b637e902f4a839",
    "height": 1280,
    "id": "7jYdNZ4DNorBpF9ZkqI6",
    "label": [],
    "labels": [],
    "name": "hollywood5.jpg",
    "numSteps": 5,
    "owner": "g2f4T0d2HBu3j7JK4iSE",
    "preprocessing": [
        "auto-orient",
        "resize:[\"Stretch to\",1280,1280]",
        "saturation:[-6]",
        "brightness:[-1]",
        "noise:[0.5,51184]"
    ],
    "preprocessingParsed": [
        {
            "name": "Auto-Orient",
            "value": "Applied"
        },
        {
            "name": "Resize",
            "value": "Stretch to 1280x1280"
        },
        {
            "name": "Saturation",
            "value": "-6%"
        },
        {
            "name": "Brightness",
            "value": "-1%"
        },
        {
            "name": "Noise",
            "value": "0.5%"
        }
    ],
    "source": "7jYdNZ4DNorBpF9ZkqI6",
    "split": "train",
    "split.niAVe8eMYrNWkWN1y3FJ.2": "train",
    "status": "generated",
    "transforms": "[\n    \"auto-orient\",\n    \"resize:[\\\"Stretch to\\\",1280,1280]\",\n    \"saturation:[-6]\",\n    \"brightness:[-1]\",\n    \"noise:[0.5,51184]\"\n]",
    "updated": {
        "_seconds": 1646738448,
        "_nanoseconds": 503000000
    },
    "updatedDate": "Mar 8, 2022",
    "updatedTime": "12:20PM",
    "updatedTimezone": "+01:00",
    "versions": [
        "niAVe8eMYrNWkWN1y3FJ/2"
    ],
    "width": 1280
}
{
    "boxes": [
        {
            "label": "Wally",
            "x": 1251,
            "y": 803,
            "width": 19,
            "height": 35
        },
        {
            "label": "Wilma",
            "x": 1248.5,
            "y": 149,
            "width": 25.5,
            "height": 6
        }
    ],
    "height": 1280,
    "key": "hollywood5_jpg.rf.23f88c72e1e33e389ee011508cfb2398.jpg",
    "width": 1280
}

Data Splits

Train	Validation	Test
2811 (80%)	351 (10%)	352 (10%)

Dataset Creation

Source Data

This dataset is a Roboflow project called "Wally v3 Computer Vision Project". You can find references below: Reference: Wally, “Wally v3 dataset.” https://universe.roboflow.com/wally/wally-v3 , mar 2022. visited on 2023-09-20.

BibTex: @misc{ wally-v3_dataset,
    title = { Wally v3 Dataset },
    type = { Open Source Dataset },
    author = { Wally },
    howpublished = { \url{ https://universe.roboflow.com/wally/wally-v3 } },
    url = { https://universe.roboflow.com/wally/wally-v3 },
    journal = { Roboflow Universe },
    publisher = { Roboflow },
    year = { 2022 },
    month = { mar },
    note = { visited on 2023-09-20 },
}

Annotations

Annotation process

Manual annotation of none, one or more annotations per snapshot saved under the "boxes" field.

Who are the annotators?

Roboflow user: Wally (https://universe.roboflow.com/wally)

Considerations for Using the Data

Social Impact of Dataset

Recreational Purposes

Additional Information

Dataset Curators

Roboflow user: Wally (https://universe.roboflow.com/wally)