Search

Digital image manipulation has become increasingly accessible and realistic with the advent of generative AI technologies. Recent developments allow for text-guided inpainting, making sophisticated image edits possible with minimal effort. This poses new challenges for digital media forensics. For example, diffusion model-based approaches could either splice the inpainted region into the original image, or regenerate the entire image. In the latter case, traditional image forgery localization (IFL) methods typically fail. This paper introduces the Text-Guided Inpainting Forgery (TGIF) dataset, a comprehensive collection of images designed to support the training and evaluation of image forgery localization and synthetic image detection (SID) methods. The TGIF dataset includes approximately 80k forged images, originating from popular open-source and commercial methods; SD2, SDXL, and Adobe Firefly. Using this data, we benchmark several state-of-the-art IFL and SID methods. Whereas traditional IFL methods can detect spliced images, they fail to detect regenerated inpainted images. Moreover, traditional SID may detect the regenerated inpainted images to be fake, but cannot localize the inpainted area. Finally, both types of methods fail when exposed to stronger compression, while they are less robust to modern compression algorithms, such as WEBP. As such, this work demonstrates the inefficiency of state-of-the-art detectors on local manipulations performed by modern generative approaches, and aspires to help with the development of more capable IFL and SID methods. The dataset can be downloaded at https://github.com/IDLabMedia/tgif-dataset., Comment: 6 pages, submitted to conference

Digital watermarking enables protection against copyright infringement of images. Although existing methods embed watermarks imperceptibly and demonstrate robustness against attacks, they typically lack resilience against geometric transformations. Therefore, this paper proposes a new watermarking method that is robust against geometric attacks. The proposed method is based on the existing HiDDeN architecture that uses deep learning for watermark encoding and decoding. We add new noise layers to this architecture, namely for a differentiable JPEG estimation, rotation, rescaling, translation, shearing and mirroring. We demonstrate that our method outperforms the state of the art when it comes to geometric robustness. In conclusion, the proposed method can be used to protect images when viewed on consumers' devices., Comment: Accepted and presented at IEEE International Conference on Consumer Electronics (ICCE) 2024

Virtual Reality, and Extended Reality in general, connect the physical body with the virtual world. Movement of our body translates to interactions with this virtual world. Only by moving our head will we see a different perspective. By doing so, the physical restrictions of our body's movement restrict our capabilities virtually. By modelling the capabilities of human movement, render engines can get useful information to pre-cache visual texture information or immersive light information. Such pre-caching becomes vital due to ever increasing realism in virtual environments. This work is the first work to predict the volume in which the head will be positioned in the future based on a data-driven binned-ellipsoid technique. The proposed technique can reduce a 1m3 volume to a size of 10cm3 with negligible accuracy loss. This volume then provides the render engine with the necessary information to pre-cache visual data.

Manipulated images are a threat to consumers worldwide, when they are used to spread disinformation. Therefore, Comprint enables forgery detection by utilizing JPEG-compression fingerprints. This paper evaluates the impact of the training set on Comprint's performance. Most interestingly, we found that including images compressed with low quality factors during training does not have a significant effect on the accuracy, whereas incorporating recompression boosts the robustness. As such, consumers can use Comprint on their smartphones to verify the authenticity of images., Comment: Will be presented at the International Conference on Consumer Electronics (ICCE) 2023 in Las Vegas, NV, USA

Manipulation tools that realistically edit images are widely available, making it easy for anyone to create and spread misinformation. In an attempt to fight fake news, forgery detection and localization methods were designed. However, existing methods struggle to accurately reveal manipulations found in images on the internet, i.e., in the wild. That is because the type of forgery is typically unknown, in addition to the tampering traces being damaged by recompression. This paper presents Comprint, a novel forgery detection and localization method based on the compression fingerprint or comprint. It is trained on pristine data only, providing generalization to detect different types of manipulation. Additionally, we propose a fusion of Comprint with the state-of-the-art Noiseprint, which utilizes a complementary camera model fingerprint. We carry out an extensive experimental analysis and demonstrate that Comprint has a high level of accuracy on five evaluation datasets that represent a wide range of manipulation types, mimicking in-the-wild circumstances. Most notably, the proposed fusion significantly outperforms state-of-the-art reference methods. As such, Comprint and the fusion Comprint+Noiseprint represent a promising forensics tool to analyze in-the-wild tampered images., Comment: Presented at the Workshop on MultiMedia FORensics in the WILD 2022, held in conjunction with the International Conference on Pattern Recognition (ICPR) 2022

In six-degrees-of-freedom light-field (LF) experiences, the viewer's freedom is limited by the extent to which the plenoptic function was sampled. Existing LF datasets represent only small portions of the plenoptic function, such that they either cover a small volume, or they have limited field of view. Therefore, we propose a new LF image dataset "SILVR" that allows for six-degrees-of-freedom navigation in much larger volumes while maintaining full panoramic field of view. We rendered three different virtual scenes in various configurations, where the number of views ranges from 642 to 2226. One of these scenes (called Zen Garden) is a novel scene, and is made publicly available. We chose to position the virtual cameras closely together in large cuboid and spherical organisations ($2.2m^3$ to $48m^3$), equipped with 180{\deg} fish-eye lenses. Every view is rendered to a color image and depth map of 2048px $\times$ 2048px. Additionally, we present the software used to automate the multi-view rendering process, as well as a lens-reprojection tool that converts between images with panoramic or fish-eye projection to a standard rectilinear (i.e., perspective) projection. Finally, we demonstrate how the proposed dataset and software can be used to evaluate LF coding/rendering techniques(in this case for training NeRFs with instant-ngp). As such, we provide the first publicly-available LF dataset for large volumes of light with full panoramic field of view, Comment: In 13th ACM Multimedia Systems Conference (MMSys '22), June 14-17, 2022, Athlone, Ireland. ACM, New York, NY, USA, 6 pages

The Art and Science Interaction Lab (ASIL) is a unique, highly flexible and modular interaction science research facility to effectively bring, analyse and test experiences and interactions in mixed virtual/augmented contexts as well as to conduct research on next-gen immersive technologies. It brings together the expertise and creativity of engineers, performers, designers and scientists creating solutions and experiences shaping the lives of people. The lab is equipped with state-of-the-art visual, auditory and user-tracking equipment, fully synchronized and connected to a central backend. This synchronization allows for highly accurate multi-sensor measurements and analysis.

Nowadays, an enormous amount of videos are streamed every day to countless users, all using different devices and networks. These videos must be adapted in order to provide users with the most suitable video representation based on their device properties and current network conditions. However, the two most common techniques for video adaptation, simulcast and transcoding, represent two extremes. The former offers excellent scalability, but requires a large amount of storage, while the latter has a small storage cost, but is not scalable to many users due to the additional computing cost per requested representation. As a third, in-between approach, network-distributed video coding (NDVC) was proposed within the Moving Picture Experts Group (MPEG). The aim of NDVC is to reduce the storage cost compared to simulcast, while retaining a smaller computing cost compared to transcoding. By exploring the proposed techniques for NDVC, we show the workings of this third option for video providers to deliver their contents to their clients.

In this paper we assess the impact of head movement on user's visual acuity and their quality perception of impaired images. There are physical limitations on the amount of visual information a person can perceive and physical limitations regarding the speed at which our body, and as a consequence our head, can explore a scene. In these limitations lie fundamental solutions for the communication of multimedia systems. As such, subjects were asked to evaluate the perceptual quality of static images presented on a TV screen while their head was in a dynamic (moving) state. The idea is potentially applicable to virtual reality applications and therefore, we also measured the image quality perception of each subject on a head mounted display. Experiments show the significant decrease in visual acuity and quality perception when the user's head is not static, and give an indication on how much the quality can be reduced without the user noticing any impairments., Comment: 6 pages, 5 figures