Domain Adaptation with Morphologic Segmentation (ADP3, CVPR 2021) – Conference Presentation
Follow @csgKAUST on X (Twitter) ... https://twitter.com/csgkaust
#DomainAdaptation with Morphologic #Segmentation
Workshop #AutonomousDriving: Perception, Prediction and Planning (ADP3)
IEEE Conference on Computer Vision and Pattern Recognition (#CVPR 2021)
Jonathan Klein¹, Sören Pirk², Dominik L. Michels¹
¹#KAUST Visual Computing Center
²#Google AI
Project Page: https://jonathank.de/research/paper/eps/index.html
Abstract: We present a novel domain adaptation framework that uses morphologic segmentation to translate images from arbitrary input domains (real and synthetic) into a uniform output domain. Our framework is based on an established image-to-image translation pipeline that allows us to first transform the input image into a generalized representation that encodes morphology and semantics - the edge-plus-segmentation map (EPS) - which is then transformed into an output domain. Images transformed into the output domain are photo-realistic and free of artifacts that are commonly present across different real (e.g. lens flare, motion blur, etc.) and synthetic (e.g. unrealistic textures, simplified geometry, etc.) data sets. Our goal is to establish a preprocessing step that unifies data from multiple sources into a common representation that facilitates training downstream tasks in computer vision. This way, neural networks for existing tasks can be trained on a larger variety of training data, while they are also less affected by overfitting to specific data sets. We showcase the effectiveness of our approach by qualitatively and quantitatively evaluating our method on four data sets of simulated and real data of urban scenes.
Видео Domain Adaptation with Morphologic Segmentation (ADP3, CVPR 2021) – Conference Presentation канала KAUST Computational Sciences
#DomainAdaptation with Morphologic #Segmentation
Workshop #AutonomousDriving: Perception, Prediction and Planning (ADP3)
IEEE Conference on Computer Vision and Pattern Recognition (#CVPR 2021)
Jonathan Klein¹, Sören Pirk², Dominik L. Michels¹
¹#KAUST Visual Computing Center
²#Google AI
Project Page: https://jonathank.de/research/paper/eps/index.html
Abstract: We present a novel domain adaptation framework that uses morphologic segmentation to translate images from arbitrary input domains (real and synthetic) into a uniform output domain. Our framework is based on an established image-to-image translation pipeline that allows us to first transform the input image into a generalized representation that encodes morphology and semantics - the edge-plus-segmentation map (EPS) - which is then transformed into an output domain. Images transformed into the output domain are photo-realistic and free of artifacts that are commonly present across different real (e.g. lens flare, motion blur, etc.) and synthetic (e.g. unrealistic textures, simplified geometry, etc.) data sets. Our goal is to establish a preprocessing step that unifies data from multiple sources into a common representation that facilitates training downstream tasks in computer vision. This way, neural networks for existing tasks can be trained on a larger variety of training data, while they are also less affected by overfitting to specific data sets. We showcase the effectiveness of our approach by qualitatively and quantitatively evaluating our method on four data sets of simulated and real data of urban scenes.
Видео Domain Adaptation with Morphologic Segmentation (ADP3, CVPR 2021) – Conference Presentation канала KAUST Computational Sciences
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![GazeboPlants: Simulating Plant-Robot Interaction with Cosserat Rods (ICRA 2024)](https://i.ytimg.com/vi/r8Q31w4bNbs/default.jpg)
![DCGrid: An Adaptive Grid Structure ... (I3D 2022) – Conference Presentation](https://i.ytimg.com/vi/4vY6RMQsBs0/default.jpg)
![Scintilla: Simulating Combustible Vegetation for Wildfires (SIGGRAPH 2024)](https://i.ytimg.com/vi/54_wO_WEOo4/default.jpg)
![Rhizomorph: The Coordinated Function of Shoots and Roots (SIGGRAPH 2023)](https://i.ytimg.com/vi/dYnO0Yqek8M/default.jpg)
![A Current Loop Model for the Fast Simulation of Ferrofluids (TVCG 2022)](https://i.ytimg.com/vi/eTRGbVpJsrw/default.jpg)
![Cyclogenesis: Simulating Hurricanes and Tornadoes (SIGGRAPH 2024)](https://i.ytimg.com/vi/ORav-_G4_Ws/default.jpg)
![Surface-Only Ferrofluids (SIGGRAPH Asia 2020)](https://i.ytimg.com/vi/MH0bMsUBZG8/default.jpg)
![Wind Erosion: Shape Modifications by Interactive Particle-based Erosion and Deposition (SCA 2020)](https://i.ytimg.com/vi/ncqphx83OBU/default.jpg)
![A Physically-inspired Approach to the Simulation of Plant Wilting (SIGGRAPH Asia 2023) – Video Talk](https://i.ytimg.com/vi/74WfKJNaKsM/default.jpg)
![DCGrid: An Adaptive Grid Structure for Memory-Constrained Fluid Simulation on the GPU (I3D 2022)](https://i.ytimg.com/vi/2E4lFfGQ3kc/default.jpg)
![Cyclogenesis: Simulating Hurricanes and Tornadoes (SIGGRAPH 2024) – Bonus Footage](https://i.ytimg.com/vi/f-G98Wz8NRA/default.jpg)
![Synthetic Silviculture: Multi-scale Modeling of Plant Ecosystems (SIGGRAPH 2019)](https://i.ytimg.com/vi/mfBt_A6YHmo/default.jpg)
![UrbanFlow: Designing Comfortable Outdoor Areas (SimAUD 2023)](https://i.ytimg.com/vi/FVNx3KzM13o/default.jpg)
![A Lennard-Jones Layer for Distribution Normalization](https://i.ytimg.com/vi/XlLpASZ2QE4/default.jpg)
![Accurately Solving Rod Dynamics with Graph Learning (NeurIPS 2021)](https://i.ytimg.com/vi/L-xS4RPvP9Y/default.jpg)
![Fire in Paradise: Mesoscale Simulation of Wildfires (SIGGRAPH 2021)](https://i.ytimg.com/vi/YE6l3IaRzsM/default.jpg)
![Ecoclimates: Climate-Response Modeling of Vegetation (SIGGRAPH 2022)](https://i.ytimg.com/vi/3BFf7mJNSQE/default.jpg)
![Weatherscapes: Nowcasting Heat Transfer and Water Continuity (SIGGRAPH Asia 2021)](https://i.ytimg.com/vi/2RzpcffFr_Q/default.jpg)
![Flexible Kokotsakis Meshes with Skew Faces: Generalization of the Orthodiag. Inv. Type (CAD 2024)](https://i.ytimg.com/vi/48H7SdvB5Ps/default.jpg)
![Interactive Wood Fracture (SCA 2020)](https://i.ytimg.com/vi/fvjIh70OsCI/default.jpg)