In between the last OpenCV AI Competition 2021 update and this one, we made a big decision- the Competition deadlines have been extended due to the ongoing COVID-19 pandemic and its effects on many teams abilities to complete their projects. The official webpage has been updated with the new dates. Along with our partners, OpenCV also recently put out a new press release focusing on the cool, new, solutions that these teams are creating over this summer.
In this post we’re featuring a short question and answer session with some of the teams who have posted videos and pics online as part of OpenCV AI Competition 2021 using the #OAK2021 hashtag. Many thanks to teams African Eagles and Calcutta Devs for agreeing to be interviewed this time! If you’d like to see your team here, drop me a line at phil @ opencv.org or hit me up on LinkedIn. If you missed the previous editions of this series featuring teams and cool videos from the Competition, go back and read Part 1, Part 2, Part 3 and Part 4! Now, onto the newest highlights!
The Latest #OAK2021 Highlights
It’s been great as the competition goes on to see more and more teams posting data, videos, and pictures for the first time! Hopefully this will accelerate even more as we enter the last leg of the Phase 2 period. To all the teams who have shared their progress: thank you!
- Team Roc4t’s project can now track dozens of cows out in the field
- Sofia, from Team Calcutta Devs, is a new Yoga app which connects OAK-D to your smart phone to help you do poses correctly and safely
- Agri Bot, the 3D printed quadruped, identifying, walking toward and watering plants
- Team Kauda demonstrated reinforcement learning with 16 agents in Unity to assist their robot arm
- Team Kauda also showed us how their robot arm can assist in worker safety and collaboration
- Team African Eagles project detects ripe and rotten fruit to intelligently classify food quality
- PalPose, a project for detecting bad posture and helping correct it
- Team EgyptIris released a dataset of Egyptian currency to the public
- Team UCB Beavers & UCB Squirrels produced a Spanish-language “Hola Mundo” or “Hello World” video for OAK-D
- Spectacular AI shows a reconstructed 3D mesh of an old WW1-era fortification with their visual-inertial SLAM technology with an AR overlay
Q&A With Team Calcutta Devs
What is your project? Briefly describe your problem statement and proposed solution
Maintaining correct posture while performing a yoga asana is important, wrong alignment of body parts may lead to injuries, and cause abnormal wear and tear of joints in the body. Public yoga classes may not help to improve the alignments and if you are more privacy concerned hiring a personal yoga trainer is quite expensive. To solve this problem, we introduce Sofia – a personal yoga instructor integrated with OAK-D for real-time pose recognition, and a voice assistant for giving proper feedback to improve posture.
Does your team have a funny “origin story?” How did you get together?
We met in our sophomore year. Souvik is a mobile app developer and IoT enthusiast. Soumi loves computer vision and deep learning. We were talking one day and suddenly came across this idea. We saw that nobody had actually implemented yoga pose classification. The first part of this process, joint points detection, was done but not the classification. We decided to take up the challenge. Soumi decided to work on the pose recognition model and get it running on OAK-D. Souvik created the app and connected it using IoT devices. You can literally control your OAK-D from your phone now!
How did you decide what problem to solve?
We had been searching for quite some time, for a project idea where we can truly show the power of a mobile app integrated with artificial intelligence. Then once we came across the problem of maintaining a correct posture in yoga sessions, and we didn’t find any user-friendly solution on the marketplace that actually solves this issue. This was the time when we started working on an AI-powered personal yoga instructor that will interact with the user to provide feedback in real-time.
What is the most exciting part of #OAK2021 to you?
We got to tinker with a lot of stuff that we didn’t have any idea about before this hackathon. Apart from that, the weekly webinar sessions are really nice. And, definitely the team updates from all over the world.
What do you think / feel upon learning you were selected for Phase 2?
We were super excited to be selected for the second phase! Couldn’t wait for the OAK-D to arrive and try out cool stuff we could do with it. We were also grateful to have been chosen from such a large and talented pool of participants.
What, if anything, has surprised you so far about the competition?
It’s amazing to see how people are contributing to open source by putting some of their codes on Github. Also, everyone in the community is so helpful and supportive. The support channels on discord have been a great help, and have accelerated our application development process considerably.
Do you have any words for your fellow competitors?
We are so inspired by the innovative applications that our competitors are creating. As per us, competition is only about winning, learning and working on solutions that can help the people and society is the primary goal. We wish them all the best for the competition.
Where should readers follow you, to best keep up with your progress? (Twitter, LinkedIn, etc)
You will get updates about the project on our LinkedIn pages:
The project source code is available on GitHub
Q&A with Team African Eagles
What is your project? Briefly describe your problem statement and proposed solution
Our project is to build an easy to use Computer Vision-based food quality classification system using the OAK-D to assist food suppliers and consumers on appropriate food and waste management. Our solution is to leverage the image-based functionalities of the OAK-D along with standard deep learning techniques such as object detection to train models that perform food quality classification.
Does your team have a funny “origin story?” How did you get together?
Fatima-Ezzahra couldn’t apply to the competition last year. So when the opportunity came, she made a post on LinkedIn looking for teammates. Ikechukwu-Nigel, who was a fellow student in a Udacity ML Scholarship program by Microsoft indicated interest. Sara, an ML engineer from the Google Community got to know about the post and shared it on her Facebook Developer Group where Abderrahim showed interest and joined the team.
How did you decide what problem to solve?
We had a series of meetings online and we had to explore different ideas of what to work on. At some point we had to even take a vote on top ideas which helped us agree on the Food quality classification project.
What is the most exciting part of #OAK2021 to you?
Getting to work together with experienced teammates, trying out several ideas with the OAK-D device, overcoming our disagreements easily as a team and helping one another makes it really exciting for all of us.
What do you think / feel upon learning you were selected for Phase 2?
It was an amazing feeling. It made us see that there is really no harm in trying, and we felt motivated to work harder in order to turn our proposal into reality. We were also excited to use the OAK-D as we had wanted before to test the power of deep learning with smart cameras.
What, if anything, has surprised you so far about the competition?
The other projects our fellow competitors are working on are inspiring and show the amazing things we could do with the OAK-D device.
Do you have any words for your fellow competitors?
I’ll say “what is worth doing is worth doing well!” It’s a privilege for us to be working in the field of AI at this present time, so let us keep learning, keep building and creating change in our various communities with what we have learnt. It always pays off!
Where should readers follow you, to best keep up with your progress? (Twitter, LinkedIn, etc)
You can find us on Linkedin and Github:
You can also find Team African Eagles on Github
More To Come!
Thank you for reading this fifth post in our series of team profiles and highlights. These are just a few of the over 250 teams participating in the competition- we wish them all the very best of luck! If you’re an AI creator who wants to join in on the fun, why not buy yourself an OAK-D from The OpenCV Store?
Stay tuned for another update soon, and follow the #OAK2021 tag on Twitter and LinkedIn for real-time updates from these great teams. Don’t forget to sign up for the OpenCV Newsletter to be notified when new posts go live, and get exclusive discounts and offers from our partners.
If you’re in the Competition and would like to see your team highlighted or interviewed here, reach out to Phil on LinkedIn! See you next time!
