Hey there, dev.to community! 🎉
We're back with a research update on our GitHub + DEV 2023 Hackathon project, ChickenVision! In case you don't know, ChickenVision is an entertaining mobile app that uses augmented reality (AR) to replace people's heads in a crowd with a chicken helmet. Today, we'll be sharing our findings on YOLOv7 pose estimation.
YOLOv7 Pose Estimation 📈
We've explored YOLOv7 Pose Estimation software for our project, and the results are amazing! 😄 YOLOv7 is self-hostable, open-source, and provides remarkable accuracy on various recordings. Here's a quick video example of how well it works:
📹 UNIC Graduation / YOLOv7 Pose Estimation Test.mp4
Challenges 🧗♂️
Despite the promising results, we're facing a few challenges:
- Overlapping masks calculations. 🎭
- Determining the position, rotation, and size of the mask. 📐
- Processing large files asynchronously. ⏳
To overcome these obstacles, we plan to investigate alternative software for image processing and pose estimation.
Stay tuned for more updates on ChickenVision and feel free to contribute to our project! Don't forget to visit our GitHub organization projectPavoculus for more information.
If you have any questions or feedback, please leave a comment or reach out to us. Let's have some fun with ChickenVision! 🐔🚀
Happy coding,
The Pavoculus team 🦃
Top comments (0)