Updating driver of LCD (MB1166-A09) for STM32 H747I-DISCO board

[mrinalinibarik]

Hi, 

I am currently working with the STM32 H747I-DISCO board, which uses the MB1166-A09 LCD daughter board. I attempted to update the driver by following the instructions from this resource: https://wiki.st.com/stm32mcu/wiki/AI:How_to_perform_people_counting_using_FP-AI-VISION1_and_STM32H747XI#Driver_update_-28required_only_when_using_MB1166_version_A09-29  

Despite following the steps, I am still experiencing a noisy display when running the people counting project in STM32CubeIDE. Could someone please provide guidance on how to resolve this issue?

[Julian E.]

Hello @mrinalinibarik ,

 

This function pack is quite old. You should consider to try using ST Model Zoo.

https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main?tab=readme-ov-file 

 

In the first read me, you will find multiple use case, but here, you are interested by object detection:

With Model zoo, you can bring your own model and data, train it, quantize it, benchmark it etc. 

In your case, I advice you to take a already trained model from the list for this particular use case and deploy it.

 

You can see that there are multiple kinds of models, with multiples input size and ready for deployment target.

For example, if you take the SSD MobileNet v2 fpn lite 0.35, you can see that in the column "target for deployment" you have your DISCO board and Camera.

If you click on the name of the model, you get a lot of information about the model performance on one or multiple data set and in particular this:

We can see that on the COCO Person dataset (which is the same use case as in the function pack) you have different versions of the same model with different input size and performance (AP here).

If you click on one of them, you will be able to find it in the repository and you can download it.

 

Now to deploy it, we will follow the readme of object detection.

When you click on object detection, you will find the object detection specific readme:

In your case, you simply want to deploy a model, not train it, quantize it etc, so you want to click on the example: deployment, chain_qd

You want to follow this tutorial.

 

You will probably need to follow first the before you start to install model zoo.

 

Good luck,

Julian