AI on STM32 : Computer Vision made simple with FP-AI-VISION1 and STM32Cube.AI

Re create step by step using : bit.ly/stm32mc...
And read out the "Jupyter notebook" used for this very application : bit.ly/stm32ai
FP-AI-VISION1 is composed of software components generated by the X-CUBE-AI Expansion Package complemented with application software components dedicated to the AI-based computer vision application.
Find out more information : www.st.com/en/...

[@manualibay7586]

This guy should start his own RUclips channel for sure

[@ooo-yt7bu]

I need this tutorial! Great👍

[@Michalszczepanski91]

great tutorial!

[@stmicroelectronics]

Glad you think so!

[@gregdavis9376]

Awesome!

[@julienpornet6597]

Hey,
By using your MCU instead of Raspberry Py, the speed of the processing will be improved ? What is the bus which communicates with the camera ? The MCU contains an os ?
Thank you for your attention.
Julien

[@jakobflocke134]

Hey, it will not be a speed upgrade comoared to the Raspberry Pi, but it will be a cost upgrade, because an stm32h7 development board (not fron st) can be baught with only 15€.
Im not sure, but the bus should be dcmi (Digital Camera Intercace) as that is an stm standard
If you mean an user os as ob most soc boards like raspberry Pi 4 / 5, then defenetely not, there is not enough Power. If you're thinking about an RTOS, then it is easily possible to set this up in STM32CUBE with just a few clicks.
Sorry for all the typos, I'm not native and not used to the english keyboard 😅

[@kavorka8855]

The link to the package is broken, cannot be downloaded :(

[@jibui5826]

Can I use STM32F407 Disco to process image ?

[@kangu2man]

Hi, I used STM32F429 Disco for image classification, but i used jpeg file instead of camera. Check my video
ruclips.net/video/WcCLhHORZiY/видео.html

[@user-pg5gg9rx1b]

Why do the screen shows a mess after i import the'STM32H747I-DISCO_Person_MobileNetv2_Quant8_IntFps_IntFlash_V310.bin' on my STM32H747I board?

[@user-ft8nj8mi2s]

How can this project be ported to F7?

[@kiruthikaganesan3137]

Whether the lcd works even after flashing the code? Kindly reply sir!!!!!

[@stmicroelectronics]

Hello. We've reached out to our colleagues about your question. Here is their answer. We hope it helps.
Yes it woks. The tutorial presented here demonstrates a way to modify a full-featured application from FP-AI-VISION1 to a different use case. i.e. from the PersonDetection classifier to a pasta (wiki article) or STM32 board (video) classifier. The LCD display, camera capture and onboard validation features will remain the same.
The modified application can be seen in action with working LCD display at 10:21 in the video or at the end of the wiki article: wiki.st.com/stm32mcu/wiki/AI:How_to_use_transfer_learning_to_perform_image_classification_on_STM32

[@kiruthikaganesan3137]

Thanks for your response. Our LCD was not working at that time . we followed all the steps as mentioned in the video but the switching option was not enabled in the camera app as mentioned in the video 4.40 , we also checked our device manager but the webcam option was not enabled.
Kindly help us rectify this issue or suggest some websites/ tutorial to capture the images through B-CAMS-OMV which will help us in proceeding our project.