The frame rate is about 3.5 sec per image (RPi overclocked to 1950 MHz)
Special made for a bare Raspberry Pi see Q-engineering deep learning examples
Paper: https://openaccess.thecvf.com/content_ICCV_2019/papers/Bolya_YOLACT_Real-Time_Instance_Segmentation_ICCV_2019_paper.pdf
| Model | size | objects | mAP | RPi 4 64-OS 1950 MHz |
|---|---|---|---|---|
| YoloV5n | 640x640 nano | 80 | 28.0 | 1.4 - 2.0 FPS |
| YoloV5s | 640x640 small | 80 | 37.4 | 1.0 FPS |
| YoloV5l | 640x640 large | 80 | 49.0 | 0.25 FPS |
| YoloV5x | 640x640 x-large | 80 | 50.7 | 0.15 FPS |
| Yoact | 550x550 | 80 | 28.2 | 0.28 FPS |
To run the application, you have to:
- A raspberry Pi 4 with a 32 or 64-bit operating system. It can be the Raspberry 64-bit OS, or Ubuntu 18.04 / 20.04. Install 64-bit OS
- The Tencent ncnn framework installed. Install ncnn
- OpenCV 64 bit installed. Install OpenCV 4.3
- Code::Blocks installed. (
$ sudo apt-get install codeblocks)
To extract and run the network in Code::Blocks
$ mkdir MyDir
$ cd MyDir
$ wget https://github.com/Qengineering/Yolact-ncnn/archive/refs/heads/master.zip
$ unzip -j master.zip
Remove master.zip and README.md as they are no longer needed.
$ rm master.zip
$ rm README.md
Your MyDir folder must now look like this:
dog.jpg
elephant.jpeg
girafe.jpeg
mumbai.jpg
onyx.jpeg
result_elephant.png
result_zebra.png
Yolact.cpb
yolact.cpp
yolact.bin (download this file from Gdrive )
yolact.param
Run Yolact.cpb with Code::Blocks.
For more info follow the instructions at Hands-On.
Many thanks to nihui again!
