VRIC: Vehicle Re-Identificaton in Context

Aytac Kanaci         Xiatian Zhu         Shaogang Gong

Computer Vision Group, School of Electronic Engineering and Computer Science, Queen Mary University of London

Summary

Example vehicle bounding-box and whole scene images of the VRIC benchmark.

(a) Samples of vehicle bounding-box images.

(b) The near and far views in a wide-view traffic scene.

(c) UA-DETRAC video scenes with different illumination due to changing weather conditions (sunny, cloudy and rainy) and time (day and night).

(d) Vehicle matching pairs (each column) from some example test vehicle instances.