挑战性图像篡改检测(CIMD)数据集
收藏资源简介:
挑战性图像篡改检测(CIMD)数据集由纽约州立大学奥尔巴尼分校计算机科学系创建,包含两个子集,分别针对基于图像编辑和基于压缩的篡改检测方法。数据集包含800张高质量的手动采集和篡改的图像,确保了样本和标注的高质量。CIMD数据集旨在为评估最先进的图像篡改检测模型提供一个可靠和准确的基准,特别关注于检测小区域篡改和相同质量因子的双重压缩情况。该数据集的应用领域包括数字取证和媒体安全,旨在解决图像篡改检测中的挑战性问题。
Challenging Image Manipulation Detection (CIMD) dataset was developed by the Department of Computer Science, University at Albany, State University of New York. It includes two subsets targeting image editing-based and compression-based tampering detection approaches respectively. The dataset contains 800 high-quality, manually captured and manipulated images, guaranteeing the high quality of both the samples and their corresponding annotations. The CIMD dataset is designed to serve as a reliable and accurate benchmark for evaluating state-of-the-art image tampering detection models, with special emphasis on detecting small-area manipulations and double compression scenarios with the same quality factor. Focused on addressing challenging issues in image tampering detection, this dataset has applications in the fields of digital forensics and media security.

- 1A New Benchmark and Model for Challenging Image Manipulation Detection纽约州立大学奥尔巴尼分校计算机科学系 · 2024年



