Stay ahead of deepfake threats with DuckDuckGoose's latest deepfake dataset, designed to enhance your identity verification systems. Benchmark your detection capabilities against industry standards and ensure your security measures are unbeatable*
.png)
Rigorously evaluate and enhance your identity verification systems, ensuring your company is well-equipped to detect and counteract deepfakes effectively.
.png)
.png)
.png)
.png)
.png)
.png)
.png){kind=icon}
Work with high-quality, real-world data, sourced from the VCD dataset, Unsplash, and Wikimedia Commons, to ensure authenticity in your testing.
.png){kind=icon}
Gain access to a meticulously curated dataset of 2,000+ samples, combining both genuine and deepfake images, offering a realistic testing ground for your detection algorithms.
.png){kind=icon}
.png){kind=icon}
A combination of genuine and deepfake images, designed specifically to help assess the effectiveness of identity verification systems in detecting and mitigating deepfake attacks.
A robust collection of 2000 images, with 850 real images (42.5%) and 1150 deepfake images (57.5%), providing a balanced dataset for thorough evaluation.
Our dataset includes various types of deepfakes: 282 face-swap samples, 598 face-swap with super-resolution, and 270 full image synthesis samples, offering a wide range of challenges for identity verification systems.
Real images in the dataset are meticulously sampled from the VCD dataset, Unsplash, and Wikimedia Commons, ensuring high-quality and diverse representations.
The dataset comes with binary labels. Users can submit their results to receive detailed performance statistics, including insights into specific deepfake types and potential biases.
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.
.svg)
.svg)
.png)
.png){kind=icon}