TensorFlow Lite Segmentation running at 4.0 FPS on bare Raspberry Pi 4
A fast C++ implementation of TensorFlow Lite Unet on a bare Raspberry Pi 4. Once overclocked to 1900 MHz, the app runs at 4.0 FPS!
https://arxiv.org/abs/1606.00915
Training set: VOC2017
Size: 257x257
Frame rate Unet Lite : 4.0 FPS (RPi 4 @ 1900 MHz - 32 bits OS)
Frame rate Unet Lite : 7.2 FPS (RPi 4 @ 1850 MHz - 64 bits OS)
Special made for a bare Raspberry Pi see: https://qengineering.eu/install-tensorflow-2-lite-on-raspberry-pi-4.html
To extract and run the network in Code::Blocks
$ mkdir MyDir
$ cd MyDir
$ wget https://github.com/Qengineering/TensorFlow_Lite_Segmentation_RPi_32/archive/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:
cat.jpg.mp4
deeplabv3_257_mv_gpu.tflite
TestUnet.cpb
Unet.cpp
Run TestUnet.cpb with Code::Blocks. Remember, you also need a working OpenCV 4 on your Raspberry.
Preferably use our installation: https://qengineering.eu/install-opencv-4.3-on-raspberry-pi-4.html
