Get Started
Home
Topics
Search
Library
Research questionHow can object detectors produce trustworthy confidence scores in degraded or ambiguous scenes?Object detectors often report confidence scores without revealing whether those scores are supported by the visual evidence. Blur, occlusion, and low texture can make this reliability difficult to judge when detections guide later decisions.
AI
Computer Vision
Evaluation & Benchmarks
Image & Video Processing
Machine Learning
Multimodal Models
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.YOLO with Kolmogorov-Arnold networks and vision-language foundation models for interpretable object detection with trustworthy multimodal AI in computer vision perceptionThe source studies vehicle and general object detection using YOLOv10 outputs on COCO and University of Bath campus images. It uses a post-hoc Kolmogorov–Arnold network surrogate based on seven geometric and semantic features to model confidence trustworthiness, with BLIP-generated scene captions providing a multimodal interface. Evidence covers degraded conditions including blur, occlusion, and low texture; broader deployment reliability is not established.research paper · Sep 4, 2026
Related questions
How can subjective image-quality assessments resist adversarial ratings that distort mean opinion scores?How can existing camera–LiDAR BEV detectors stay reliable under domain shifts and sensor failures without retraining?How can LLMs produce reliable confidence estimates for deciding when to defer outputs to humans?How can RGB-D salient-object detection remain reliable when depth measurements are missing or corrupted?