Luckfox Forums

Both instances utilize the opencv-mobile driver to capture images from the sc3336 camera, process the images, and display the results on an LCD screen. The object recognition routine uses the Yolov5 convolutional neural network model, capable of recognizing 80 types of objects and displaying confidence levels.


The instance displays the results of model inference on the Pico-LCD-1.3 screen, requiring the screen to be treated as a Framebuffer device driver.

**only convert the Pico-LCD 1.3 screen to 3.5 ResTouch LCD for our facial purpose

Respected sir,
I am very grateful for your help,
I am trying to use alternative ways......
https://wiki.luckfox.com/Luckfox-Pico/RKNN-example/
Please if you could guide me with this

Hello sir,
Please refer to the description in the wiki.
If there is a problem with the source program or article description, you can send a separate question request.
If modifications or features need to be added, we can discuss them on the forum.
During discussions, we can learn from each other.
Under normal circumstances, if you need to add or modify functions, we will not directly provide the program.
After all, the technical team’s human resources are limited, so please understand.

Sir,
For face recognition the fnorm parameter is coming same for every different face detected.
so is there any other parameter for face recognition?
if not then how can it be improve?

I’ve recently come across the Luckfox Pico Max with face recognition and I’m really impressed by how compact yet powerful it is for embedded AI projects. The accuracy and speed of the facial detection surprised me, especially considering its size and price point. If you’re curious about the technology behind it or want to create something similar, this article on how to build a face recognition system offers a great breakdown of the process and tools involved. Definitely worth checking out if you're diving into computer vision projects!