Study indicators without promising certainty.
A NaXhio.Tech research line on visual traces associated with synthetic images and methods for analyzing authenticity.
Image authenticity
A visual reading of the proposal
- What the person brings
Images with a known origin
- The proposed work
Examine signs and methods
- What allows work to continue
Compare findings and limits
The proposal
An image can look convincing and still require a critical examination of its origin. This research line explores which visual indicators can be studied and under which conditions they provide useful information about generated images.
The public work focuses on studying visual artifacts, analyzing synthetic images and evaluating methods. It is intended for researchers and teams interested in understanding the scope and limitations of these tools.
Scope and evolution
The priority is to formulate measurable questions and compare methods against an evaluation set. An indicator does not by itself amount to proof of authenticity: it matters to know when a method is correct, when it is wrong and when it cannot reach a conclusion.
Planned capabilities
Study of visual artifacts
Synthetic image analysis
Evaluation of authenticity methods
At this stage
Open research line. Results and their application require evaluation.
Next horizon
Define an evaluation set and compare methods with explicit limits.