Professorship of Data Science at VISAPP 2025
17 December 2024
From February 26 to 28, 2025, the 20th International Conference on Computer Vision Theory and Application (VISAPP 2025) took place in Porto, Portugal. Amon Soares de Souza represented the professorship and presented his research results.
In his lecture, Amon Soares de Souza presented results obtained as part of the Deepfake project. Among other things, the project deals with so-called adversarial attacks in image classification. Such attacks manipulate images so subtly that a model is forced to output an incorrect classification. These images are then particularly susceptible to transformations through post-processing such as compression. In this context, a JPEG compression invariance for adversarial optimization was introduced.
The International Conference on Computer Vision Theory and Applications aims to become a major focal point for researchers, engineers and practitioners in the field of computer vision methods, systems and applications. Several tracks cover all different aspects of computer vision, including image and video processing and analysis; image and video understanding; mobile and egocentric vision for humans and robots as well as applications and services.
Figure 2: Impression of the poster session.
Image sources: Amon Soares de Souza (RI CODE) and Wikimedia Commons / Joergsam / License CC BY-SA 3.0
