Need Advice on How to Start Contributing to Computer Vision Exercises in Robotics Academy

[Im-Arth1307]

Good Evening Everyone,
I wanted to ask for your advice on how I can start contributing towards the Computer Vision Exercises available on the Robotics Academy. I've already successfully setup and launched Robotics Academy using Docker and I have also successfully completed the Line Follow Exercise and some parts of the Basic Computer Vision Exercise. It would be of huge help if anyone could guide me as to how I can proceed and make meaningful contributions.
Thank you
Regards,
Atharva Jakhetiya

[jmplaza]

Hi @Im-Arth1307 ,

one idea for volunteer contribution is to make an object tracking application, based on GoodFeaturesToTrack and on OpticalFlow functions at OpenCV, just using the Basic Computer Vision exercise. It should include some object segmentation and object tracking.

Cheers,

JoseMaria

[Im-Arth1307]

Thank you so much for the guidance @jmplaza sir. I've started researching the concept based on your suggestion and will begin working on the object tracking application soon. Once it's ready, I’ll open a PR for review.
Best Regards,
Atharva Jakhetiya

[Im-Arth1307]

Good Evening @jmplaza Sir,
I have been working on the task you gave me and was able to successfully implement an object tracker withing the basic computer vision exercise, operating as a user. The output showcases a virtual bounding box to identify the object of interest, with its movement tracked using nodes marked on it. It would be of huge help if you could guide me as to how to proceed from here.
Should I focus on updating the existing documentation to incorporate the object tracking feature, similar to how grayscale, color filter, edge filter etc are documented for the Basic Computer Vision exercise ? Or would it be more appropriate to create a separate exercise, similar to how the "Follow Face" exercise is organized?
Thank you
Best Regards,
Atharva Jakhetiya

[jmplaza]

Hi @Im-Arth1307 ,

I don't give tasks, just suggestions of collaboration on this open source project :-)

Feel free to share here the video of your object tracking solution. Creating a new specific exercise is not planned, this could be for now an extension inside Basic Computer Vision exercise. We are recovering the DeepLearning based object detection (and tracking exercise). More opinions and info is still needed before deciding a new exercise.

Best,

[Im-Arth1307]

Good Evening @jmplaza Sir,
Thanks for the clarification!
I understand that a new specific exercise is not planned at the moment.

I have uploaded the video demonstrating my object tracking solution on YouTube. Here’s the link: YouTube link

Looking forward to any feedback or suggestions for improving.

Best Regards,
Atharva Jakhetiya

[jmplaza]

Nice video @Im-Arth1307 .

Regarding this, @lu164 is also working on introducing a topic (subsexercise?) inside the BasicComputerVision exercise about the use of some OpenCV methods such as GoodFeaturesToTrack and tracking. Maybe you can betatest her PRs or create your own one.

In addition, there is a short term goal you can contribute to: extending BasicComputerVision exercise to get frames from YouTube videos. You are welcome in case you want to explore tentative solutions.

[Im-Arth1307]

Thank you @jmplaza Sir,
I’ll look into @lu164’s contributions and assist with testing. I’ll also start exploring some possible approaches and see what could work best for getting frames from YouTube videos.
Best Regards,
Atharva Jakhetiya