New benchmark evaluates vehicle attribute classification in surveillance

Researchers have introduced the Unconstrained Vehicle Identification Benchmark (UVIB) to evaluate vehicle attribute classification in diverse surveillance scenarios. This benchmark, comprising 84,835 images from seven Brazilian datasets, addresses the common issue of model performance degradation when transitioning from controlled environments to real-world surveillance due to variations in viewpoint, occlusion, and lighting. The study evaluated four architectures—EfficientNetV2-S, ResNet-50, ViT/B-16, and YOLO11s-cls—demonstrating that domain shift significantly impacts performance, particularly for vehicle make and model recognition suitability and color clarity, more so than the choice of architecture. AI

IMPACT This benchmark could lead to more robust vehicle attribute classification models for intelligent transportation systems.

RANK_REASON The item describes a new benchmark and evaluation of computer vision models for a specific task, fitting the research category. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• Brazilian Portuguese
• EfficientNetV2-S
• intelligent transportation system
• ResNet-50
• Unconstrained Vehicle Identification Benchmark
• YOLO11s-cls

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →