cancel
Showing results for 
Search instead for 
Did you mean: 

Face Recognition MCU Selection

koksoybedirhan
Associate III

Hello,

I have been conducting research for a while for a new project initiation. The project aims to design a smart access control system for 500 users using STM32. In designing this, we are implementing face recognition with FP_AI_FACEREC in the STM32H7 series and creating the display design with touchGFX. I have made progress in the project, and by adjusting the RAM settings of the face recognition library, I was able to accommodate 500 users. However, the library doesn't work at the desired accuracy level, especially when similar users are enrolled after reaching 50 users. It sometimes mistakenly identifies a user and allows access. This has raised doubts about the use of the H7 series, and I am considering selecting an MP135 series card for the processor. If someone is knowledgeable about this, could they provide information about the performance of the H7 and MP135 series for AI? Can we use higher accuracy models if we switch to MP135, or would we achieve similar performance as with the H7 series? I can provide more detailed information if needed for a more specific answer.

Best regards.

1 ACCEPTED SOLUTION

Accepted Solutions
TDK
Guru

I don't think that the accuracy of the algorithm is going to be better or worse on any particular platform due to that platform. Surely it would have more to do with the resources available and processing speed, and quality of the algorithm. The MPUs generally have much more processing capability than the MCUs and would therefore be a more capable choice.

If you feel a post has answered your question, please click "Accept as Solution".

View solution in original post

5 REPLIES 5
TDK
Guru

I don't think that the accuracy of the algorithm is going to be better or worse on any particular platform due to that platform. Surely it would have more to do with the resources available and processing speed, and quality of the algorithm. The MPUs generally have much more processing capability than the MCUs and would therefore be a more capable choice.

If you feel a post has answered your question, please click "Accept as Solution".

Fix the algorithm. Determine what parameters you can change related to training, and re-training

As you have all the imagery and process setup, and understood, you'd probably be in the best position to try the software on both platforms to determine if one is materially better than the other, or has speed or memory advantages.

Tips, buy me a coffee, or three.. PayPal Venmo Up vote any posts that you find helpful, it shows what's working..

Hi, as you said, it will be the best option to try both of them, Thank you.

I will try both and repost the result again.

VRICH.1
ST Employee

Hi,

The FP-AI-FACEREC1 is a demonstration package for evaluation purpose only according to its license terms. The aim is to enable the feasibility of a face recognition application on STM32H7 series.

Typically, it is recommended for a max of 20 to 50 enrolled faces.

The hardware choice between STM32H7 and STM32MP1 will not affect the algorithm performances (accuracy) but the inference speed will be higher with an MP1, as explained in a previous reply.

For production grade quality algorithms, ST is working with Biometrics Partners who are experts in this field.

For contacting ST Biometrics partners please write an email to edge.ai@st.com

Best regards,

Hi @VRICH.1 

Thanks for your compherensive answer.

Best regards.