dtd

# [DTD: Describable Textures Dataset](https://www.robots.ox.ac.uk/~vgg/data/dtd/)

The Describable Textures Dataset (DTD) is an evolving collection of textural images in the wild, annotated with a series of human-centric attributes, inspired by the perceptual properties of textures.
This data is made available to the computer vision community for research purposes

## Usage

```python
from datasets import load_dataset

dataset = load_dataset('tanganke/dtd')
```

- **Features:**
- **Image**: The primary data type, which is a digital image used for classification. The format and dimensions of the images are not specified in this snippet but should be included if available.
- **Label**: A categorical feature representing the texture or pattern class of each image. The dataset includes 46 classes with descriptive names ranging from 'banded' to 'zigzagged'.
- **Class Labels**:
- '0': banded
- '1': blotchy
- '2': braided
- ...
- '45': wrinkled
- '46': zigzagged

- **Splits**: The dataset is divided into training and test subsets for model evaluation.
- **Training**: containing 3760 examples with a total size of 448,550 bytes.
- **Test**: containing 1880 examples with a total size of 220,515 bytes.

# [DTD：可描述纹理数据集（Describable Textures Dataset）](https://www.robots.ox.ac.uk/~vgg/data/dtd/)

可描述纹理数据集（Describable Textures Dataset，简称DTD）是一个持续演进的野外自然纹理图像集合，其标注基于纹理的感知特性，采用一系列以人为中心的属性完成注释。本数据集面向计算机视觉社区开放，仅用于科研用途。

## 使用方法

python
from datasets import load_dataset

dataset = load_dataset('tanganke/dtd')

- **核心特征：**
- **图像：** 作为核心数据类型，用于纹理分类任务。本代码片段未指定图像的格式与尺寸，若存在相关可用信息，应补充完整。
- **标签：** 为分类特征，用于表征每张图像对应的纹理或模式类别。本数据集共包含46个带有描述性名称的类别，命名范围从“带状（banded）”延伸至“锯齿状（zigzagged）”。
- **类别标签：**
- '0'：带状（banded）
- '1'：斑点状（blotchy）
- '2'：编织状（braided）
- ...
- '45'：褶皱状（wrinkled）
- '46'：锯齿状（zigzagged）

## 数据集划分

本数据集划分为训练集与测试集子集，用于模型性能评估。
- **训练集：** 包含3760条样本，总数据量为448,550字节。
- **测试集：** 包含1880条样本，总数据量为220,515字节。