AI summaryⓘ
The authors created a new benchmark called UVIB to test how well computer models can recognize vehicle features like orientation, occlusion, and color in real-world surveillance images, which are often messy and different from controlled test conditions. They gathered a large set of vehicle images from different sources and labeled them consistently to evaluate four popular model types. Their tests showed that changes in environment or camera setup (domain shift) hurt performance more than the choice of model architecture, especially for recognizing vehicle make/model and color clarity. The authors suggest that it is important to test models in ways that reflect real-world challenges, not just ideal cases.
Intelligent Transportation SystemsVehicle Make and Model RecognitionDomain ShiftOcclusionBenchmark DatasetVehicle OrientationColor ClarityDeep Learning ArchitecturesCross-Domain Evaluation
Authors
Sergio M. Silva, Otavio T. Remer, Gabriel E. Lima, Lucas Wojcik, Rayson Laroca, David Menotti
Abstract
Vehicle attribute analysis is a key component of Intelligent Transportation Systems (ITS), supporting applications such as vehicle identification, traffic monitoring, and forensic investigation. However, models trained under controlled conditions often degrade in real surveillance scenarios due to changes in viewpoint, occlusion, illumination, and sensor characteristics. This paper introduces Unconstrained Vehicle Identification Benchmark (UVIB), a benchmark for evaluating three operational vehicle-analysis tasks: front/rear orientation, occlusion-related suitability for Vehicle Make and Model Recognition (VMMR), and color clarity. The benchmark contains 84,835 vehicle images from seven public Brazilian datasets, grouped into surveillance and general acquisition domains, with unified binary annotations that were not jointly available in the original sources. Four representative architectures, EfficientNetV2-S, ResNet-50, ViT/B-16, and YOLO11s-cls, are evaluated under mixed-domain, cross-domain, and cross-dataset protocols. The results show that domain shift has a stronger impact than architecture choice, with substantial degradation in cross-domain settings, especially for VMMR suitability and color clarity. While orientation generalizes more reliably, VMMR suitability remains affected by class imbalance and ambiguous occlusions, and color clarity is highly sensitive to illumination and sensor modality. These findings highlight the need for benchmarks and evaluation protocols that explicitly measure operational robustness beyond standard in-domain accuracy. The proposed benchmark is publicly available at https://github.com/UFPR-IPASP-PR/uvib-vehicle-attributes/.