Tire pattern classification is an important means to provide clues for traffic accident processing. With the rapid increase in the number of vehicles, it is in urgent need to develop efficient and automatic tire pattern image classification and recognition system, so as to further improve the work efficiency of law enforcement departments. Due to the large memory requirement and large amount of computation, traditional deep learning networks cannot run on mobile devices as well as embedded devices. Therefore, it is a challenging task to design a tire pattern image classification model based on lightweight network.
More than 8000 tire pattern images, a total of 50 classes. Each class contains tire pattern surface images with different proportions and angles, and tire indentation images in practical application scenarios.
Design a lightweight network model for tire pattern image classification. The model should provide satisfactory performance in tire pattern classification with small size and fast classification speed. Specifically, the performance of the model will be evaluated from the following aspects: the number of parameters, FLOPs (floating-point operations per second) and classification accuracy. The query images tested will include both tire surface patterns and tire indentation patterns. The tire features obtained must be robust so as to demonstrate satisfied performance when tested on low-quality tire patterns taken from actual scenarios.
Parameters, FLOPs and classification accuracy.
Participants can be college students, graduate researchers or professionals;
1-6 team members and 1-2 advisers in each group;
Participants need to provide source codes, trained model, algorithm description and testing results reported in the form of GC paper(following the format requirement ).
Milestones of the challenge:
May. 20-June. 10, 2022 Registering for the competition.
June. 1-June. 10, 2022 Training Data Collection.
June. 8-Nov. 1, 2022 Team work.
Nov. 2-Nov. 10,2022 Team report preparation.
Nov. 11, 2022 Submission.
Nov. 12-Nov. 25, 2022 Testing, performance evaluation, ranking.
Nov. 26, 2022 Winner notification, Competition close.
Nov. 27-Dec. 12,2022 Prepare for winner certificate and GC session at the conference.
Dec 13-16, 2022, VCIP2022 at Suzhou, China.
Prizes: The challenge sets the following prizes,
One 1st prize, two 2nd prize, 3 third prize for 3 teams;
Winning team will receive a certification as well as prize in cash, and will be asked to present their work at the conference.
Tire pattern image data:
VCIP2022, main challenge website: http://www.vcip2022.org/index.html
Xi'an University of Posts and Telecommunications, Xi’an, China
Shandong Artificial Intelligence Institute, Shandong, China
Data provided by:
Center for Image and Information Processing, Xi'an University of Posts and Telecommunications, Xi'an, China, 710121.
Key Lab. of Electronic Information Processing for Crime Scene Investigation Applications, Ministry of Public Security, China.
For any inquiry, please email us at: email@example.com, firstname.lastname@example.org
VCIP2022 Challenge registration form
We look forward to meeting you at the Grand Challenge!