Video Data
Video data analysis is a rapidly evolving field focused on extracting meaningful information from video sequences for diverse applications. Current research emphasizes developing robust models for tasks like anomaly detection (using memory-based methods and deep learning), action recognition (leveraging CNNs, RNNs, and transformers), and video generation (with action-conditioned models), often addressing challenges related to data scarcity and computational efficiency. These advancements are crucial for improving various sectors, including autonomous driving, surveillance, healthcare, and scientific research, by enabling more accurate and efficient analysis of increasingly prevalent video data.
Papers
Synthesizing Annotated Image and Video Data Using a Rendering-Based Pipeline for Improved License Plate Recognition
Andreas Spruck, Maximilane Gruber, Anatol Maier, Denise Moussa, Jürgen Seiler, Christian Riess, André Kaup
PTSD in the Wild: A Video Database for Studying Post-Traumatic Stress Disorder Recognition in Unconstrained Environments
Moctar Abdoul Latif Sawadogo, Furkan Pala, Gurkirat Singh, Imen Selmi, Pauline Puteaux, Alice Othmani