Key Point and Vehicle Orientation Annotation for VeRi-776 dataset

====================================================================

Introduction

VeRi-776 is a large-scale benchmark dateset for vehicle Re-Id in the real-world urban surveillance scenario. It contains over 50,000 images of 776 vehicles captured by 20 cameras covering an 1.0 km^2 area in 24 hours, which makes the dataset scalable enough for vehicle Re-Id and other related research.

This repo has annotations of key point location and vehicle orientation for VeRi-776 dataset, which is used in our ICCV'17 paper Orientation Invariant Feature Embedding and Spatial Temporal Regularization for Vehicle Re-identification.

Get VeRi-776 dataset

Please refer to this repo.

Key points' definition

We defined 20 key points in a vehicle, which we think are the most discriminative locations or some main vehicle components, for instance, wheels, lamps, auto logo and so on. You can find our definition of the 20 key points in the figure and the table bellow.

index	location	index	location
1	left-front wheel	11	left rear-view mirror
2	left-back wheel	12	right rear-view mirror
3	right-front whee l	13	right-front corner of vehicle top
4	right-back wheel	14	left-front corner of vehicle top
5	right fog lamp	15	left-back corner of vehicle top
6	left fog lamp	16	right-back corner of vehicle top
7	right headlight	17	left rear lamp
8	left headlight	18	right rear lamp
9	front auto logo	19	rear auto logo
10	front license plate	20	rear license plate

Orientation's definition

We classify the orientation of a vehicle into 8 categories, according to which face(s) of the vehicle is visible in this view :

0	front
1	rear
2	left
3	left front
4	left rear
5	right
6	right front
7	right rear

Annotation file format

In each line in the annotation file, the format is:

img_path x1 y1 x2 y2 ... x20 y20 orien

(x_i,y_i) is the location of the ith key point of a vehicle, and orien is the orientation label.

Citation

If you find this repo useful in your research, please consider to cite:

@InProceedings{Wang_2017_ICCV,
	author = {Wang, Zhongdao and Tang, Luming and Liu, Xihui and Yao, Zhuliang and Yi, Shuai and Shao, Jing and Yan, Junjie and Wang, Shengjin and Li, Hongsheng and Wang, Xiaogang},
	title = {Orientation Invariant Feature Embedding and Spatial Temporal Regularization for Vehicle Re-Identification},
	booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
	month = {Oct},
	year = {2017}
}

Contact

Please concact Zhongdao Wang (wcd17@mails.tsinghua.edu.cn) if you have questions about the annotations.