This month, Adobe announced in its blog that Adobe researchers and their collaborators at UC Berkeley discovered ways to detect edits in photos made using the Face Aware Liquify feature on Photoshop. The research has a broader goal of developing methods for better detecting manipulation in image, audio, video and other types of documents.
The researcher created a training set with pictures from the internet for the Adobe AI. In experiments, the neural network tool can achieve results of detecting doctored facial images as high as 99%. The tool can also calculate and revert a manipulated image to its original state.
Popular image editing apps on mobile device and professional creative tools such as Adobe Photoshop have changed people’s standards and expectations for beauty. Unrealistic standards of beauty is made a reality on digital and print media alike and commercially packaged to boost sales in different industries.
Despite the commercial value of manipulated images, the change in our visual culture can affect people’s life in many ways. People may feel obliged to make changes to their photos before they post on social media. Expectations from the others based on the new visual standards may render people more emotionally vulnerable in terms of how they look. Drastic measures like plastic surgeries are taken to close the gap between the visual fantasy and the reality.
Adobe also realizes the ethical and cultural implications of their creative tools and acknowledges that fake content has become a pressing issue. This new research is Adobe’s way of taking responsibility of the impact of their technologies. Hopefully, it can bring back trust and authority in digital media.