Qengineering / TensorFlow_Lite_SSD_RPi_64-bits

TensorFlow Lite SSD on bare Raspberry Pi 4 with 64-bit OS at 24 FPS
https://qengineering.eu/install-ubuntu-18.04-on-raspberry-pi-4.html
BSD 3-Clause "New" or "Revised" License
39 stars 6 forks source link

Best model/approach for detect hands in RPi-4 #5

Open MyraBaba opened 1 year ago

MyraBaba commented 1 year ago

Hi,

We have a application to protect my friends hands . running on normal PC 25fps

what would be the best model and AI platform to detect hand like below to prevent accidents in RPi ? should be fast as possible as

Camera fixed up and see directly thx

Screen Shot 2022-07-25 at 17 27 57 Screen Shot 2022-07-25 at 17 28 27
MyraBaba commented 1 year ago

is this palm tflite model can work in this your program ?

https://github.com/google/mediapipe/tree/master/mediapipe/modules/palm_detection

Qengineering commented 1 year ago

The mentioned tflite model (palm_detection_lite.tflite) can work in Pose detection. After some changes to the code of course. It will run at ~13 FPS. You could improve the inference times by porting the model to int8. Another (minor) improvement can be running the software on a Rpi without desktop.

MyraBaba commented 1 year ago

Thanks for direction.

edited model and 17 to 21 :) but nothing detected and drawing :)

at least not crashes but nothing drawing to screen.

need little help or will wait your hand detection example ..

Best

MyraBaba commented 1 year ago

@Qengineering Would you mind to show me directin which parts need to be changed ?

MyraBaba commented 1 year ago

I assume https://github.com/google/mediapipe/tree/master/mediapipe/modules/palm_detection with https://github.com/google/mediapipe/blob/master/mediapipe/modules/hand_landmark/hand_landmark_full.tflite

work together to draw fingers

found this also :

https://github.com/FeiGeChuanShu/ncnn_nanodet_hand

Qengineering commented 1 year ago

Please see https://github.com/Qengineering/Hand-Pose-ncnn-Raspberry-Pi-4