A new research paper explores methods for improving surveillance systems by analyzing dynamic scenes, object detection, and motion tracking. The study compares various motion triggers and deep learning detectors, finding that detector-based triggers store significantly fewer idle frames than traditional motion triggers. The research also highlights that increasing the input resolution for YOLO11n models substantially improves detection accuracy, and INT8 quantization can reduce latency without a significant drop in performance. However, the study notes that tracking accuracy is often limited by detection failures rather than association issues. AI
IMPACT This research could lead to more efficient and accurate surveillance systems by optimizing detection and tracking algorithms.
RANK_REASON Research paper published on arXiv detailing AI methods for surveillance. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- CNN
- Google Summer of Code 2018
- k-nearest neighbors algorithm
- MOG2
- Shahran Rahman Alve
- VisDrone
- YOLO11n
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
