BiometricLab-DTU-Splice Dataset
收藏资源简介:
BiometricLab-DTU-Splice Dataset是由德里理工大学信息科技系生物识别研究实验室创建的新型图像拼接检测数据集,包含自动生成和手动编辑两种变体,共计22524个样本。该数据集通过Python代码和Adobe Photoshop软件生成,每个样本包括原始图像、拼接图像及其对应的二值掩码,适用于深度学习模型提取区分性特征而不会过拟合。数据集主要用于图像拼接检测研究,旨在解决图像篡改检测问题,提高社会对视觉数据的信任度。
The BiometricLab-DTU-Splice Dataset is a novel image splicing detection dataset created by the Biometrics Research Laboratory, Department of Information Technology, Delhi Technological University. It contains two variants: automatically generated and manually edited, with a total of 22,524 samples. Generated via Python code and Adobe Photoshop software, each sample includes the original image, the spliced image, and its corresponding binary mask, which is suitable for deep learning models to extract discriminative features without overfitting. The dataset is primarily used for image splicing detection research, aiming to address the problem of image tampering detection and enhance public trust in visual data.




