Hello everyone,
The Aircraft Engine Defect Detector app uses Google’s Gemini API to automatically identify and annotate defects in aircraft engine images, generating detailed inspection reports for maintenance and regulatory compliance.
Video demo can be found via this link:
Kindly let me know your thoughts. I’m happy to receive feedback.
Many thanks,
John Uzoma
AI is known to hallucinate, so it would have been useful to show examples on how it performs on non-defective parts, see if there are false negatives. And also defect types besides scratches. The image quality isn’t very high, so it’s not clear whether those really are scratches or just lighting.
Also I think it would score some more points if you could make the problem a little clearer - who needs this? What are they using now? Is it faster to take photos than have a trained eye?
I have a friend who built an aeronautics startup. Basically the aircraft engineers have a very, very expensive toolbox to prevent mistakes like leaving tools inside the engine. The startup idea was to automate photos of the toolbox to identify that all the tools were in place. But back then, the image recognition tech was not good enough and it was too expensive to build something like this. So if you wanted to turn this into a startup, that’s one thing you could look at.
thank you so much for your feedback. will take those into consideration.