SemilleroCV/BWMP2
收藏Hugging Face2024-09-02 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/SemilleroCV/BWMP2
下载链接
链接失效反馈官方服务:
资源简介:
---
task_categories:
- image-classification
language:
- en
tags:
- Images
pretty_name: 'Material Classification Hands On '
size_categories:
- n<1K
dataset_info:
config_name: plain_text
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': Brick
'1': Metal
'2': Paper
'3': Plastic
'4': Wood
splits:
- name: train
num_examples: 120
- name: test
num_examples: 30
license: mit
---
# Dataset Card for Material Classification
## Dataset Description
- **Homepage:** https://semillerocv.github.io/proyectos.html
- **Repository:** https://github.com/Sneider-exe/Clasificacion_Materiales
### Dataset Summary
The Material_classification_2U dataset consists of 150 256x256 color images, categorized into 5 classes with 30 images per class. The dataset is divided into two main subsets: 120 images for training and 30 images for testing. Each image is labeled into one of the following five categories: Brick, Metal, Paper, Plastic, and Wood.
### Supported Tasks and Leaderboards
- `image-classification`: The goal of this task is to classify a given image into one of 5 classes.
### Languages
English
## Dataset Structure
### Data Instances
A sample from the training set is provided below:
```
{
'image': <PIL.PngImagePlugin.PngImageFile image mode=RGB size=256x256>,
'label': 1
}
```
### Data Fields
- image: A `PIL.Image.Image` object containing the 256x256 image. Note that when accessing the image column: `dataset['train']["image"]` the image file is automatically decoded. Decoding of a large number of image files might take a significant amount of time.
- label: 0-4 with the following correspondence
'0': Brick
'1': Metal
'2': Paper
'3': Plastic
'4': Wood
### Data Splits
The dataset is divided into two main subsets: Train and Test.
- **Train Split:**
- **Number of Images:** 120
- **Distribution:** 24 images per class
- **Test Split:**
- **Number of Images:** 30
- **Distribution:** 6 images per class
Both splits are stratified, ensuring that each class is proportionally represented in both the Train and Test subsets. This means that the percentage of images for each class remains consistent across both splits, providing a balanced and representative distribution for model training and evaluation.
### Citation Information
```
@TECHREPORT{
author = {Brayan Sneider Sánchez, Dana Meliza Villamizar, Cesar Vanegas, Juan Jose Calderón},
title = {BMWP2},
institution = {Universidad Industrial de Santander},
year = {2024}
}
```
task_categories:
- 图像分类(image-classification)
language:
- 英语(en)
tags:
- 图像(Images)
pretty_name: '实操材料分类(Material Classification Hands On)'
size_categories:
- n<1K
dataset_info:
config_name: 纯文本(plain_text)
features:
- name: 图像(image)
dtype: 图像
- name: 标签(label)
dtype:
class_label:
names:
'0': 砖块(Brick)
'1': 金属(Metal)
'2': 纸张(Paper)
'3': 塑料(Plastic)
'4': 木材(Wood)
splits:
- name: 训练集(train)
num_examples: 120
- name: 测试集(test)
num_examples: 30
license: MIT许可证(mit)
---
# 材料分类数据集卡片
## 数据集说明
- **主页:** https://semillerocv.github.io/proyectos.html
- **代码仓库:** https://github.com/Sneider-exe/Clasificacion_Materiales
### 数据集概览
本Material_classification_2U数据集共包含150张256×256像素的彩色图像,划分为5个类别,每类含30张图像。数据集分为两个主要子集:120张用于训练,30张用于测试。每张图像的标签对应以下五类之一:砖块、金属、纸张、塑料以及木材。
### 支持任务与排行榜
- `图像分类(image-classification)`:该任务的目标是将输入图像分类至5个类别之一。
### 语言
英语
## 数据集结构
### 数据样例
以下展示一条训练集样例:
{
'image': <PIL.PngImagePlugin.PngImageFile image mode=RGB size=256x256>,
'label': 1
}
### 数据字段
- 图像(image):`PIL.Image.Image` 对象,存储256×256像素的图像。请注意,当访问图像列时(如 `dataset['train']["image"]`),图像文件会自动解码。解码大量图像文件可能会消耗较多时间。
- 标签(label):取值范围为0至4,对应关系如下:
'0': 砖块(Brick)
'1': 金属(Metal)
'2': 纸张(Paper)
'3': 塑料(Plastic)
'4': 木材(Wood)
### 数据拆分
数据集分为训练集与测试集两个主要子集。
- **训练集拆分:**
- **图像数量:** 120
- **类别分布:** 每类24张图像
- **测试集拆分:**
- **图像数量:** 30
- **类别分布:** 每类6张图像
两个拆分均采用分层抽样,确保每个类别在训练集与测试集中的占比一致。这意味着每个类别的图像占比在两个子集间保持统一,为模型训练与评估提供了均衡且具有代表性的数据分布。
### 引用信息
@TECHREPORT{
author = {Brayan Sneider Sánchez, Dana Meliza Villamizar, Cesar Vanegas, Juan Jose Calderón},
title = {BMWP2},
institution = {Universidad Industrial de Santander},
year = {2024}
}
提供机构:
SemilleroCV



