A5081937579- Advanced Vision and Imaging
- Image Enhancement Techniques
- Image Processing Techniques and Applications
- Underwater Acoustics Research
- Computer Graphics and Visualization Techniques
- Advanced SAR Imaging Techniques
- Explainable Artificial Intelligence (XAI)
- Autonomous Vehicle Technology and Safety
- Advanced Image Processing Techniques
- Advanced Image and Video Retrieval Techniques
- Neural Networks and Applications
- Machine Learning and Data Classification
- Radar Systems and Signal Processing
- Urban Heat Island Mitigation
- Radiative Heat Transfer Studies
- Optical measurement and interference techniques
- 3D Modeling in Geospatial Applications
- Target Tracking and Data Fusion in Sensor Networks
Mercedes-Benz (Germany)
2022-2024
Vision in adverse weather conditions, whether it be snow, rain, or fog is challenging. In these scenarios, scattering and attenuation severly degrades image quality. Handling such inclement however, essential to operate autonomous vehicles, drones robotic applications where human performance impeded the most. A large body of work explores removing weather-induced degradations with dehazing methods. Most methods rely on single images as input struggle generalize from synthetic...
We propose Gated Stereo, a high-resolution and long-range depth estimation technique that operates on active gated stereo images. Using high dynamic range passive captures, Stereo exploits multi-view cues alongside time-of-flight intensity from gating. To this end, we method with monocular prediction branch which are combined in final fusion stage. Each block is supervised through combination of self-supervision losses. facilitate training validation, acquire synchronized dataset for...
The simulation of rare edge cases such as adverse weather conditions is the enabler for deployment next generation autonomous drones and vehicles into where human operation error-prone. Therefore, settings must be simulated accurately possible computationally efficient, so to allow training deep learning algorithms scene understanding, which require large-scale datasets disallowing extensive Monte Carlo simulations. One computationally-expensive step light sources in scattering media, can...
Neural fields have been broadly investigated as scene representations for the reproduction and novel generation of diverse outdoor scenes, including those autonomous vehicles robots must handle. While successful approaches RGB LiDAR data exist, neural reconstruction methods radar a sensing modality largely unexplored. Operating at millimeter wavelengths, sensors are robust to scattering in fog rain, and, such, offer complementary active passive optical techniques. Moreover, existing highly...
Current autonomous driving technologies are being rolled out in geo-fenced areas with well-defined operation conditions such as time of operation, area, weather and road conditions. In this way, challenging adverse weather, slippery or densely-populated city centers can be excluded. order to lift the restriction allow a more dynamic availability functions, it is necessary for vehicle autonomously perform an environment condition assessment real identify when system cannot operate safely...
No augmented application is possible without animated humanoid avatars. At the same time, generating human replicas from real-world monocular hand-held or robotic sensor setups challenging due to limited availability of views. Previous work showed feasibility virtual avatars but required presence 360 degree views targeted subject. To address this issue, we propose HINT, a NeRF-based algorithm able learn detailed and complete model viewing angles. We achieve by introducing symmetry prior,...
Reconstructing outdoor 3D scenes from temporal observations is a challenge that recent work on neural fields has offered new avenue for. However, existing methods recover scene properties, such as geometry, appearance, or radiance, solely RGB captures often fail when handling poorly-lit texture-deficient regions. Similarly, recovering with scanning LiDAR sensors also difficult due to their low angular sampling rate which makes expansive real-world difficult. Tackling these gaps, we introduce...
Vision in adverse weather conditions, whether it be snow, rain, or fog is challenging. In these scenarios, scattering and attenuation severly degrades image quality. Handling such inclement however, essential to operate autonomous vehicles, drones robotic applications where human performance impeded the most. A large body of work explores removing weather-induced degradations with dehazing methods. Most methods rely on single images as input struggle generalize from synthetic...
We propose Gated Stereo, a high-resolution and long-range depth estimation technique that operates on active gated stereo images. Using high dynamic range passive captures, Stereo exploits multi-view cues alongside time-of-flight intensity from gating. To this end, we method with monocular prediction branch which are combined in final fusion stage. Each block is supervised through combination of self-supervision losses. facilitate training validation, acquire synchronized dataset for...