yizhangdev/pascal-voc

---
license: apache-2.0
task_categories:
- object-detection
- image-segmentation
- image-classification
---
## Pascal VOC
#### Dataset Summary
The Pascal Visual Object Classes (VOC) dataset is a widely used benchmark in the field of computer vision. It is designed for object detection, image classification, semantic segmentation, and action classification tasks. The dataset provides a comprehensive set of annotated images covering 20 object classes, allowing researchers to evaluate and compare the performance of various algorithms.
**Note**: This dataset repository contains all editions of PASCAL-VOC, each file is identified with the year.

#### Dataset Structure
**Images:** The dataset contains 178k images.
**Annotations:** Annotations include object bounding boxes, object class labels, segmentation masks, and action labels.
**Classes:** 20 object classes: person, bicycle, car, motorbike, aeroplane, bus, train, boat, bird, cat, dog, horse, sheep, cow, elephant, bear, zebra, giraffe, and potted plant.
**Supported Tasks**
**Image Classification:** Assigning a label to an image from a fixed set of categories.
**Object Detection:** Identifying objects within an image and drawing bounding boxes around them.
**Semantic Segmentation:** Assigning a class label to each pixel in the image.
**Action Classification:** Identifying the action being performed in the image.

#### Applications
The Pascal VOC dataset is used for:

- Benchmarking and evaluating computer vision algorithms.
- Training models for image classification, object detection, and segmentation tasks.

#### Data Collection and Annotation

**Data Sources**
The images were collected from Flickr and other sources, ensuring a diverse and representative sample of real-world scenes.

A**nnotation Process**
Annotations were carried out by a team of human annotators. Each image is labeled with:

- Bounding boxes for object detection.
- Class labels for each object.
- Pixel-wise segmentation masks for semantic segmentation.
- Action labels indicating the action performed by the objects in the image.

#### License
The Pascal VOC dataset is released under the Creative Commons Attribution 2.5 License. Users are free to share, adapt, and use the dataset, provided appropriate credit is given.

#### Citation
If you use the Pascal VOC dataset in your research, please cite the following paper:

```
@article{Everingham10,
author = {Mark Everingham and
Luc Gool and
Christopher K. I. Williams and
John Winn and
Andrew Zisserman},
title = {The Pascal Visual Object Classes (VOC) Challenge},
journal = {International Journal of Computer Vision},
volume = {88},
number = {2},
year = {2010},
pages = {303-338},
}