Noise is a fundamental problem in digital photography, as it is inherit to the imaging process,and it degrades image quality. State-of-the-art digital cameras, including smartphones, include a denoising algorithm that removes noise to re-store a high quality image. Your task in this challenge is to develop a denoising method that reduces image noise, whilst preserving image detail.
November 1, 2018
Release of training dataset via confirmation email.
November 30, 2018
Submission deadline. Your team must make its submission by this deadline.
December 12, 2018
Award Ceremony at British Museum.
The AI Challenge is a programming competition organized by Huawei. You pick your team to work together, and we pick a Challenge in the field of Artificial Intelligence for you to solve. Are you up for the challenge?
The Terms & Conditions for Huawei AI Challenge gather all rules that apply to the organizers and participants of the competition. You can find a summary of the most important ones below:
This competition is open to undergraduate, post-graduate, and research (PhD) student participants in UK universities. Please note that only email address ending with *ac.uk* will be accepted in the competition. You are welcome to consult with more senior researchers (e.g. Professors, post-docs) but the work should be completed by students. At most three participants are permitted on a team. An individual can only be on one team
How to enter
The individual Participant or, if a team is entering, a member of that team as its representative on behalf of the others, must register with the Organiser through http://eucompetition.huawei.com/uk/ . The Organiser will then send via email to the *ac.uk* email address provided at registration an email which will include: (i) further instructions and (ii) a dataset consisting of a number of images (the ‘Organiser’s Material’). The Participants must also inform the organiser via email email@example.com when they are ready for submission. The organiser will then assign a login account for the Evaluation System to the participants so that they could gain access to the competition and submit a competition entry. Only login with our assigned account will be valid for the competition.
For the competition, teams will submit solutions through an Evaluation System (a web application provided by Codalab). You will need to sign in the Evaluation System using the login information that the organiser assigned to you. The Evaluation System will also feature a live scoreboard so that you’ll be able to see how the other teams perform. Maximum 10 entries a day is allowed in the competition. Participants may submit their best performing solution to the leaderboard. Only solutions submitted to the leaderboard will be evaluated for prizes.
The Organizer aims to motivate research of image de-noising technology. The Organizer will provide each Participant with a training database consisting of 975 images and a test database consisting of 30 evaluation images (“Organizer’s material”). Participants are to use their own method(s) using the Organizer’s material to ensure their method(s) is able to de-noise Huawei’s material and produce 30 de-noised images as a result.
Participants shall submit their de-noised images to the Evaluation System. They can use any programming language / framework they want. The organiser will ask participants to provide their code and will verify the valid submissions for the top performing teams.
We will award cash prizes to the three teams with top performance and the top three teams will have the opportunity to visit the Huawei Headquarters in China. Every participant will also get a certificate of qualification to the competition during the Award Ceremony
We expects all Participants to be respectful for the rules of the Competition and towards other Participants. In particular, Participants should not take any action that would prevent other Participants from taking part in the Competition in good conditions. Any cheating by using commercial tools, applying it to the noisy images, and submitting the de-noised images as their submission is not allowed for the competition.