Challenge: share your ISPU neural network implementation!

[Denise SANFILIPPO]

Dear community members,
Here’s your chance to win some prizes from ST until November 29, 2024!

You can join the challenge by sharing your ISPU neural network implementation. We’ll provide instructions and hints during the ISPU webinar held on October 29, 2024.

• https://content.st.com/iot-applications-with-neural-networks-in-mems-sensors.html

Your proposal must relate to embedding a neural network on an ST MEMS sensor with ISPU.
The ISPU is an ultralow power, computationally efficient, high-performance programmable core that can execute signal processing and AI algorithms in the edge.

It is available in the LSM6DSO16IS and ISM330IS sensors from STMicroelectronics.

Simply create a comment in this thread. You can also attach files to document your solution (.log, .mp4, .docx, .jpg, .png, .xlsx, .7z, .hex, .c, .cpp, .zip).

Use the “kudos” button on ideas or implementations from other members to indicate that you like them. Everybody can vote on implementations and ideas, and the most voted one will win the prize.

The following guidelines can help you come up with implementations and ideas to share in this thread:

• Implement a new neural network model for ISPU (for instance starting from examples available on GitHub and model zoo).
  
  
• Try to optimize and deploy the neural network model on ISPU with MEMS Studio, ST Edge AI Core, or ST Edge AI Developer Cloud.
  
  
• Share performance results (accuracy, model size, code size, model parameters, inference time, etc.) of the created model and ISPU configuration.

We will review your solutions on November 29, 2024 and send a PM to the winner of the challenge to get your information regarding shipment.

The challenge ends November 29, 2024, but your ideas and implementations are of course always welcome, even after the challenge!

[Lorenzo BRACCO]

The ST Edge AI Core for ISPU tutorial has been updated with support for the new X-NUCLEO-IKS4A1 motion MEMS and environmental sensor expansion board, check it out!

[Hank_Xiao]

I developed two demos using the LSM6DSV16X sensor and MEMS Studio software, demonstrating how to combine the STM32H503CBT6 MCU and LPS28DFW barometer to perform smart gesture recognition and sports equipment recognition.

1. Demo 1: Smart Pen Gesture Recognition
   Recognizing Steady, Idle, Writing, and Other gestures. The system can differentiate various writing postures in real-time.

2. Demo 2: Badminton Racket Recognition
   Recognizing three states: Idle, Grip, and Swing, suitable for sports monitoring and smart sports equipment.

To share my implementation experience more effectively, I have written the following blogs and recorded corresponding videos:

Blog1：https://blog.csdn.net/qq_24312945/article/details/142692648

Video1：https://www.bilibili.com/video/BV14V45eyEh9/

Blog2：https://blog.csdn.net/qq_24312945/article/details/142692810

Video2：https://www.bilibili.com/video/BV1Ba45eaEKM

Blog3：https://blog.csdn.net/qq_24312945/article/details/142705360

Video3：https://www.bilibili.com/video/BV1624VeAEAG/

Blog4：https://blog.csdn.net/qq_24312945/article/details/142705468

Video4：https://www.bilibili.com/video/BV1Kb4FeME9R/

 While promoting ST products, I have encountered some challenges and would appreciate additional support and resources. If you have any collaboration opportunities or suggestions, feel free to contact me:
Email: a845656974@outlook.com

[Denise SANFILIPPO]

Hi @Hank_Xiao ,

please post this comment inside the MEMS&Sensors community forum.
The content of your post can be useful to other users of that forum and it is out of scope for this challenge. In fact here ST is encouraging the development of applications based on ISPU embedded inside the MEMS sensors LSM6DSO16IS and ISM330IS.

Thanks!

[Hank_Xiao]

I just saw this challenge and haven't used the LSM6DSO16IS and ISM330IS yet, but I feel they are quite similar.