PRNU-Bench
收藏资源简介:
PRNU-Bench是一个用于相机识别的基准数据集,包含从126个独特的传感器采集的12,960张图片,这些传感器来自114个设备,涵盖了智能手机、平板电脑、网络摄像头、视频相机和无反光镜相机等多个类别。数据集通过在特定环境中从不同的视角采集图片,确保了图片内容的多样性,并且所有图片都是在自然场景下拍摄的,以便更好地模拟现实世界的使用情况。该数据集还包含2021年至2024年间发布的最新设备型号,为相机识别提供了更具挑战性的评估。
PRNU-Bench is a benchmark dataset dedicated to camera identification. It comprises 12,960 images collected from 126 unique sensors across 114 devices, covering various categories including smartphones, tablets, webcams, video cameras and mirrorless cameras. The dataset ensures the diversity of image content by capturing images from different perspectives in a specific environment, and all images are taken in natural scenes to better simulate real-world usage scenarios. Additionally, the dataset includes the latest device models released between 2021 and 2024, providing a more challenging evaluation benchmark for camera identification tasks.

- 1PRNU-Bench: A Novel Benchmark and Model for PRNU-Based Camera Identification布加勒斯特大学 · 2025年



