189 research outputs found
EDM-Research/DIMO_ObjectDetection: v1.0
No full textObject detection for the DIMO dataset. Uses the Mask-RCNN model. This is the official implementation of Analysis of Training Object Detection Models with Synthetic Data, published in BMVC: British Machine Vision Conference, 2022.
Source code for the following scientific publication:
Vanherle, B., Moonen, S., Van Reeth, F., and Michiels, N. (2022). Analysis of Training Object Detection Models with Synthetic Data. 33rd British Machine Vision Conference 2022, BMVC 2022, London, UK, November 21-24, 2022. Retrieved from https://bmvc2022.mpi-inf.mpg.de/0833.pdfPILS SBO: Product Inspection with Little Supervision. Flanders Make (Belgium). awardNumber:null. 02ndjfz59BOF Special Research Fund. Hasselt University. awardNumber:null. 10.13039/50110000955
Antwerpen, Romeinse stad?
No full textIs Antwerpen ooit een Romeinse stad geweest? In dit boek wordt eerst ingegaan op de stand van zaken in het archeologisch onderzoek. Daarna komen de geschreven bronnen aan bod en de belangwekkende historische besluiten die daaruit kunnen worden getrokken voor de ontstaansgeschiedenis van de stad, aansluitend bij de archeologische gegevens. Belangrijk voor het zelfbeeld van de Antwerpenaren zijn echter ook de vele legenden die door de eeuwen heen rond het ontstaan van de stad zijn geweven, vaak met politieke bedoelingen. De meest markante daarvan is wel de stelling dat het Antwerps de taal was van het Paradijs?00Met bijdragen van Jan M.F. Van Reeth, Guido Cuyt, Alfred Michiels en Toon van Hal. Jan M.F. Van Reeth is classicus en oriëntalist. Hij is onder meer voorzitter van het Nederlands Klassiek Verbond en doceert over godsdienstgeschiedenis van het Nabije Oosten
3D performance capture for facial animation
No full textThis work describes how a photogrammetry based 3D capture system can be used as an input device for animation. The 3D Dynamic Capture System is used to capture the motion of a human face, which is extracted from a sequence of 3D models captured at TV frame rate. Initially the positions of a set of landmarks on the face are extracted. These landmarks are then used to provide motion data in two different ways. First, a high level description of the movements is extracted, and these can be used as input to a procedural animation package (i.e. CreaToon). Second the landmarks can be used as registration points for a conformation process where the model to be animated is modified to match the captured model. This approach gives a new sequence of models, which have the structure of the drawn model but the movement of the captured sequence
Highly stylised drawn animation
No full textIn this paper we argue for our NPAR system as an effective 2D alternative to most of NPR research which is focused on frame coherent stylised rendering of 3D models. Our approach gives a highly stylised look to images without the support of 3D models, and yet they still behave as though animated by drawing, which they are.
First, a stylised brush tool is used to freely draw extreme poses of characters. Each character is built up of 2D drawn brush strokes which are manually grouped into layers. Each layer is assigned its place in a drawing hierarchy called a Hierarchical Display Model (HDM). Next, multiple HDMs are created for the same character, each corresponding to a specific view. A collection of HDMs essentially reintroduces some correspondence information to the 2D drawings needed for in-betweening and, in effect, eliminates the need for a true 3D model.
Once the models are composed the animator starts by defining keyframes from extreme poses in time. Next, brush stroke trajectories defined by the keyframe HDMs are in-betweened automatically across intermediate frames. Finally, each HDM of each generated in-between frame is traversed and all elements are drawn one on another from back to front.
Our techniques support highly rendered styles which are particularly difficult to animate by traditional means including the ‘airbrushed’, scraperboard, watercolour, Gouache, ‘ink-wash’, and the ‘crayon’ styles.
We believe our system offers a new fresh perspective on computer aided animation production and associated tools.
Keywords:Artist driven, stylised modelling, stylised animation, computer animation, computer-assisted animation, NPR, NPAR
High corticosterone levels in prenatally stressed rats predict persistent paradoxical sleep alterations.
No full textPrenatal stress predisposes rats to long-lasting disturbances that persist throughout adulthood (e.g. high anxiety, dysfunction of the hypothalamo-pituitary-adrenal axis, and abnormal circadian timing). These disturbances parallel to a large extent those found in depressed patients, in which hypercortisolemia and sleep alterations may be related to stress-inducing events. We studied sleep-wake parameters in control and prenatally stressed adult rats (3-4 months old) and examined possible relationships with their corticosterone levels (determined at 2 months of age). Under baseline conditions, prenatally stressed rats showed increased amounts of paradoxical sleep, positively correlated to plasma corticosterone levels. Other changes include increased sleep fragmentation, total light slow-wave sleep time, and a slight decrease in the percentage of deep slow-wave sleep relative to total sleep time. During recovery sleep from acute restraint stress, all sleep changes persisted and were correlated with stress-induced corticosterone secretion. High corticosterone levels under baseline conditions as well as an acute stress challenge may thus predict long-term sleep-wake alterations in rats. Taken together with other behavioral and hormonal abnormalities in prenatally stressed animals, the pronounced changes in sleep-wake parameters that are similar to those found in depressed patients suggest that prenatal stress may be a useful animal model of depression.Journal ArticleResearch Support, Non-U.S. Gov'tSCOPUS: ar.jinfo:eu-repo/semantics/publishe
Differential Methods in Character Rigging
No full textMost 3D animation software comes with various mesh deformers for character
rigging, based on long known and widely used algorithms, like skeletal subspace
deformation or freeform deformation. We demonstrate how these methods can
be improved upon and made more flexible while still working interactively as
part of complicated character rigs, using the same workflow as before.
We first consider the interpolation of affine transformations, and the solution
of differential equations, whose properties will be taken advantage of in the
algorithms we implemented. Next we consider the problem of shape interpolation,
and discuss a representation for mesh deformations that encodes them
in affine transformations per triangle, such that they can be interpolated and
edited in ways difficult to achieve otherwise. Then we discuss skeletal subspace
deformation and recent methods to get rid of its well known artifacts while still
being nearly as efficient. Additionally we demonstrate how t
VATr++: Choose Your Words Wisely for Handwritten Text Generation
No full textStyled Handwritten Text Generation (HTG) has received significant attention in recent years,propelled by the success of learning-based solutions employing GANs,Transformers,and,preliminarily,Diffusion Models. Despite this surge in interest,there remains a critical yet understudied aspect - the impact of the input,both visual and textual,on the HTG model training and its subsequent influence on performance. This work extends the VATr [1] Styled-HTG approach by addressing the pre-processing and training issues that it faces,which are common to many HTG models. In particular,we propose generally applicable strategies for input preparation and training regularization that allow the model to achieve better performance and generalization capabilities. Moreover,in this work,we go beyond performance optimization and address a significant hurdle in HTG research - the lack of a standardized evaluation protocol. In particular,we propose a standardization of the evaluation protocol for HTG and conduct a comprehensive benchmarking of existing approaches. By doing so,we aim to establish a foundation for fair and meaningful comparisons between HTG strategies,fostering progress in the field
Reeth, F.: Multimodal Interaction in a Collaborative Virtual Brainstorming Environment
No full textAbstract. In this paper we present our work in setting up a collaborative virtual environment (CVE) framework which is built to support collaborative creative meetings for geographically dispersed participants. Similar to real life, we rely on the use of quick drawings or sketches as a means of communication to convey new ideas, thoughts or other metadata to other individuals. Furthermore, we concentrate on facilitating the (collaborative) interaction process through the use of four modalities. The first modality is direct manipulation, which is suitable for direct interaction with the networked environment. Secondly, we look at interaction through gesturing symbols in order not to distract the user’s attention from the meeting. As a third modality we consider interaction through menu and widget manipulation. A fourth modality is established by a camera interface. We expect that the combination of the intuitive interface and the realtime visualisation of the virtual environment leads to a better understanding and realisation of one’s ideas in an early phase of the cooperation
Differential Methods in Character Rigging
No full textMost 3D animation software comes with various mesh deformers for character
rigging, based on long known and widely used algorithms, like skeletal subspace
deformation or freeform deformation. We demonstrate how these methods can
be improved upon and made more flexible while still working interactively as
part of complicated character rigs, using the same workflow as before.
We first consider the interpolation of affine transformations, and the solution
of differential equations, whose properties will be taken advantage of in the
algorithms we implemented. Next we consider the problem of shape interpolation,
and discuss a representation for mesh deformations that encodes them
in affine transformations per triangle, such that they can be interpolated and
edited in ways difficult to achieve otherwise. Then we discuss skeletal subspace
deformation and recent methods to get rid of its well known artifacts while still
being nearly as efficient. Additionally we demonstrate how t
Prenatal stress in rats predicts immobility behavior in the forced swim test. Effects of a chronic treatment with tianeptine.
No full textPrenatally-stressed (PS) rats are characterized by a general impairment of the hypothalamo-pituitary-adrenal (HPA) axis and sleep disturbances indicating that this model has face validity with some clinical features observed in a subpopulation of depressed patients. The prolonged corticosterone secretion shown by PS rats in response to stress was positively correlated with an increased immobility behavior in the forced swim test. To investigate the predictive validity of this model, a separate group of animals was chronically treated with the antidepressant tianeptine (10 mg/kg i.p. for 21 days). Such chronic treatment reduced in PS rats immobility time in the forced swim test. These findings suggest that the PS rat is an interesting animal model for the evaluation of antidepressant treatment. (C) 2003 Elsevier B.V. All rights reserved
- …
