metadata

license: apache-2.0
dataset_info:
  features:
    - name: width
      dtype: int64
    - name: height
      dtype: int64
    - name: image
      dtype: image
    - name: objects
      struct:
        - name: bbox
          sequence:
            sequence: float64
        - name: category
          sequence: string
        - name: color
          list:
            - name: alpha
              dtype: float64
            - name: blue
              dtype: float64
            - name: green
              dtype: float64
            - name: red
              dtype: float64
        - name: radius
          sequence: float64
        - name: text
          sequence: string
  splits:
    - name: train
      num_bytes: 1253458059.322
      num_examples: 7846
  download_size: 1160884066
  dataset_size: 1253458059.322
task_categories:
  - object-detection
tags:
  - ui
  - design
  - detection
size_categories:
  - n<1K

Dataset: Mobile UI Design Detection

Introduction

This dataset is designed for object detection tasks with a focus on detecting elements in mobile UI designs. The targeted objects include text, images, and groups. The dataset contains images and object detection boxes, including class labels and location information.

Dataset Content

Load the dataset and take a look at an example:

>>> from datasets import load_dataset
>>>> ds = load_dataset("mrtoy/mobile-ui-design")
>>> example = ds[0]
>>> example
{'width': 375,
 'height': 667,
 'image': <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=375x667>,
 'objects': {'bbox': [[0.0, 0.0, 375.0, 667.0],
   [0.0, 0.0, 375.0, 667.0],
   [0.0, 0.0, 375.0, 20.0],
   ...
  ],
  'category': ['text',
   'rectangle',
   'rectangle',
   ...]}}

The dataset has the following fields:

image: PIL.Image.Image object containing the image.
height: The image height.
width: The image width.
objects: A dictionary containing bounding box metadata for the objects in the image:
- bbox: The object’s bounding box (xmin,ymin,width,height).
- category: The object’s category, with possible values including rectangle、text、group、image
- color: The object’s color, text color or rectangle color, or None
- radius: The object’s color, rectangle radius, or None
- text: text content, or None

You can visualize the bboxes on the image using some internal torch utilities.

import torch
from torchvision.ops import box_convert
from torchvision.utils import draw_bounding_boxes
from torchvision.transforms.functional import pil_to_tensor, to_pil_image

item = ds[0]
boxes_xywh = torch.tensor(item['objects']['bbox'])
boxes_xyxy = box_convert(boxes_xywh, 'xywh', 'xyxy')
to_pil_image(
    draw_bounding_boxes(
        pil_to_tensor(item['image']),
        boxes_xyxy,
        labels=item['objects']['category'],
    )
)

Applications

This dataset can be used for various applications, such as:

Training and evaluating object detection models for mobile UI designs.
Identifying design patterns and trends to aid UI designers and developers in creating high-quality mobile app UIs.
Enhancing the automation process in generating UI design templates.
Improving image recognition and analysis in the field of mobile UI design.