VISIGRAPP_DC 2023 Abstracts

Full Papers
Full Papers
Paper Nr: 5
Title:

Drone-based Population Monitoring of Wildlife Using Light-Field Samples and Video Sequences in Wooded Environments Utilizing Artificial Intelligence
Authors:

Christoph Praschl
Abstract: The threat of climate catastrophe is endangering the integrity of habitats such as forests on our planet. In current forecasts by the Intergovernmental Panel on Climate Change, the habitats on our planet will become even more endangered in the coming years. Intact ecosystems such as forests, however, are the basis of our existence in a complex interplay of countless species given by the biodiversity of these systems. If individual species from these interactions are lost or become prevalent, ecosystems affected by this, threaten to become unbalanced. A crucial aspect to preserve this balance is the preservation and promotion of biodiversity. This requires the monitoring of population levels, for example in the form of area-wide observation of wildlife. However, safeguarding biodiversity can only be done with the help of accurate surveying techniques, such as counting and analyzing animal populations. This process of population monitoring is needed to detect changes in numbers as well as inter-/intragroup distributions within different species. Such changes represent an indicator of a stable population. Conversely, overpopulation (e.g., invasive species) or underpopulation (e.g., due to disease, loss of habitat, ...) can be identified, which need to be addressed by appropriate means e.g., to protect endangered species or to control the spread of invasive species. Current methods (e.g., camera traps) use random sampling to estimate the population and density. However, such methods are unsuitable for large-scale areas. In contrast, camera-based observation using uncrewed aerial vehicles (UAV) can be used over large areas, but in forested areas, area-wide observation and counting of wildlife are very limited due to dense vegetation. Accounting for this obscuring forest cover is possible, for the first time, using a new technique called airborne light-field sampling (ALFS), which allows uncovering the forest floor (e.g., wildlife). ALFS is based on light-field technology, in which a RGB and/or thermal video/image sequence of a flyover is combined with position data and an elevation model of the terrain. Unfortunately, ALFS is highly dependent on a static scenario without or with only little movement. For the use case of wild-life monitoring, this can be an issue, as animals may react to the presence of UAVs and may engage in escape behavior. Therefore, the population monitoring should be carried out in a two-folded manner: using integral images (created with ALFS) and geo-referenced video sequences.
Paper Nr: 6
Title:

Effectiveness of Virtual Reality in Learning 3D Transformations in Computer Graphics and Impact on Spatial Skills
Authors:

Maha Alobaid
Abstract: In computer graphics, three-dimensional (3D) transformations are an essential topic and are employed in the modelling, texturing, view transformations and rendering processes. 3D transformations are one fundamental concept that students find challenging, it requires a wide range of skills including programming, math, problem-solving, and spatial abilities. Despite the fact that most learning aids facilitate the teaching of 3D transformations, the cause of many problems in learning is impeded by a lack of spatial skills. The ability to interpret representations of 3D transformations and create mental images of their effects appears to be lacking in many students. The strong evidence for a correlation between spatial skills and performance in computer graphics. Computer graphics directly affect students' spatial abilities, and these abilities should be received more attention in learning computer graphics. The improvement of spatial abilities may benefit greatly from augmented and virtual reality tools. This study contributes to enhancing the learning of 3D transformations in computer graphics through virtual reality to improve spatial skills.
Paper Nr: 7
Title:

A Layered Approach to Constrain Signing Avatars
Authors:

Paritosh Sharma
Abstract: Synthesis of sign language from a formal linguistic model using data-driven animation techniques is a challenging task. The process is either expensive and repetitive or produces utterances which lack comprehension. We introduce a layer-based approach to define a signing avatar for solving posture constraints and present how it can be used for a complete data-driven multi-track sign language synthesis system. Our method synthesizes a sign language description model by combining low-level linguistic constraints as well as pre-animated actions. To unify these techniques, we formalize our posture using separate layers, which gives us control over the low-level skeleton and mesh specification. Finally, we build this system on top of an open-source animation toolkit.
Paper Nr: 9
Title:

Proposal of an Adaptive Learning System Applied in Games for Children with Down Syndrome
Authors:

Matheus Faria
Abstract: Video games, in addition to representing an extremely relevant field of entertainment and market, have been widely used as a case study in artificial intelligence for representing a problem with a high degree of complexity. In such studies, the investigation of approaches that endow player agents with the ability to retrieve relevant information from game scenes and player's action stands out, since such information can be very useful to improve their learning ability. These kind of agents can be used to enhance the players experience in adaptive games, as they will learn to model their profile and adapt game elements according to their behavior. Motivated by such facts, the present work proposes the implementation of tutorial agents apt to assist the development of the cognitive skills of individuals with some psycho-motor weakness (for example, children with Down Syndrome). The idea, in this case, is to make these agents able to perceive the vulnerabilities of such individuals by analyzing the game events of their recorded matches. From these game events, the agents will be able to abstract the actions executed by these people in various game situations and to map the specific situations in which their decision-making was fragile. From this point on, the agents' engine must try to direct the game in order to provoke the occurrence of such situations and present some clues to help the player to better deal with them (and, consequently, stimulating in the improvement of their limitations).