Robin Cole

Add custom object detection model

1

See * https://blog.roboflow.com/how-to-train-a-tensorflow-lite-object-detection-model * https://github.com/sayakpaul/E2E-Object-Detection-in-TFLite

Add Dockerfile for rpi?

2

RPi requires a slightly modified Dockerfile: ``` RUN pip3 install https://dl.google.com/coral/python/tflite_runtime-2.1.0.post1-cp37-cp37m-linux_armv7l.whl ``` Build takes a while, maybe 20 mins

Resources: * https://medium.com/analytics-vidhya/facenet-on-modile-part-3-cc6f6d5752d6 * https://machinelearningmastery.com/how-to-develop-a-face-recognition-system-using-facenet-in-keras-and-an-svm-classifier/ * https://github.com/robmarkcole/convert-tensorflow-mobilefacenet-model-to-tflite (includes .tflite file) * https://arxiv.org/abs/1804.07573

Add transfer learning example for classification

1

https://keras.io/guides/transfer_learning/ using dogs-vs-cats which can be downloaded [as here](https://keras.io/examples/vision/image_classification_from_scratch/)

Use https://realpython.com/using-flask-login-for-user-management-with-flask/

On https://www.tensorflow.org/lite/models/image_classification/overview The only notable difference is that the labels file does not include an index. Example code on https://github.com/tensorflow/examples/blob/master/lite/examples/image_classification/raspberry_pi/classify_picamera.py

Currently have to manually `docker push robmarkcole/object-detection-app` but presumably this can be automated on releases

https://github.com/ricardodeazambuja/ALPR-EdgeTPU-RPI0

https://github.com/ricardodeazambuja/Maple-Syrup-Pi-Camera

https://github.com/google-coral/aiy-maker-kit