Description:

The Penn-Fudan Database is a specialized image dataset designed for pedestrian detection. The images capture scenes from urban streets and university campuses. This dataset is widely used for research and development in computer vision, particularly in applications related to autonomous vehicles, surveillance, and urban planning. It includes detailed annotations of pedestrian locations, making it an invaluable resource for training and evaluating pedestrian detection algorithms.

Download Dataset

Details:

Subjects: Pedestrians, with at least one in each image.

Image Count: 170 images.

Pedestrian Count: 345 labeled pedestrians.

Locations: 96 images from the University of Pennsylvania, 74 from Fudan University.

Pedestrian Heights: Range from 180 to 390 pixels.

Characteristics: All pedestrians are upright.

Additional Information:

Use Case: Ideal for experiments in pedestrian detection.

Annotations: Includes detailed labeling for accurate model training.

Quality: High-resolution images ensuring clear visibility of pedestrians.

Applications:

Research: Pedestrian detection in urban environments.

Safety Systems: Enhancing the accuracy of pedestrian detection in autonomous vehicles.

Urban Planning: Analyzing pedestrian movement patterns for better infrastructure development.

This dataset is sourced from Kaggle.