A5088740469- Advanced Optical Sensing Technologies
- Anomaly Detection Techniques and Applications
- Sports Analytics and Performance
- Robotics and Sensor-Based Localization
- Video Analysis and Summarization
- Advanced Chemical Sensor Technologies
- 3D Surveying and Cultural Heritage
- Remote Sensing and LiDAR Applications
- Species Distribution and Climate Change
- Software System Performance and Reliability
- Allergic Rhinitis and Sensitization
UCLouvain
2021-2024
Tracking and identifying athletes on the pitch holds a central role in collecting essential insights from game, such as estimating total distance covered by players or understanding team tactics. This tracking identification process is crucial for reconstructing game state, defined athletes' positions identities 2D top-view of pitch, (i.e. minimap). However, state videos captured single camera challenging. It requires position viewpoint to localize identify within field. In this work, we...
Recognizing the types of pollen grains and estimating their proportion in mixture samples collected a specific geographical area is important for agricultural, medical, ecosystem research. Our paper adopts convolutional neural network automatic segmentation species microscopy images, proposes an original strategy to train such at reasonable manual annotation cost. approach founded on large dataset composed pure images. It first (semi-)manually segments foreground, i.e. grains, background...
The SoccerNet 2024 challenges represent the fourth annual video understanding organized by team. These aim to advance research across multiple themes in football, including broadcast understanding, field and player understanding. This year, encompass four vision-based tasks. (1) Ball Action Spotting, focusing on precisely localizing when which soccer actions related ball occur, (2) Dense Video Captioning, describing with natural language anchored timestamps, (3) Multi-View Foul Recognition,...
We propose a novel approach to localize 3D object from the intensity and depth information images provided by Time-of-Flight (ToF) sensor. Our method uses two CNNs. The first one raw as input, segment floor pixels, which extrinsic parameters of camera are estimated. second CNN is in charge segmenting object-of-interest. As main innovation, it exploits calibration estimated prediction represent geometric coordinate system that attached ground, thus independent elevation. In practice, both...