Extract a Model.zip , In this zip you will find frozen_inference_graph.pb a model file and labelmap.pbtxt a label file, Put this files in working directory.
Run a currency_det_im.py, if you want to detect from an image otherwise for video use currency_detection_webcam.py
Project Structure
I given my prepared model to you so you don't need to train it, just follow me:)
It can detect (10,20,50,100 Rupees) old notes, because of i am just testing & learning abou this that's why i choosen less classes,but now i am working a system which can detect all indian currency with old & new notes and even coins also,but it will take time, i am gathering a dataset for it.
I used own image dataset for this model, Images captured with mobile.
I used Tensorflow API of object detection but used own images.
I used faster_rcnn_inception_v2_coco model for training ,you can use any other model for training.
Screenshots
10 Rupees detection from live video
100 Rupees detection from live video
50 Rupees detection from image
100 Rupees detection from image
Notes
It will require high processing power(I have 8 GB RAM & 2 GB GC)
If you think it will recognise notes like human then please leave it ,it's have many bugs ,but currently i am solving it.
Quality of images matters which you are giving as input for a prediction ,please use clear images.