Gesture classifying model

Introduction

Pip package used in sign language translation system available in the following repository:

https://github.com/kacper1095/translation-system-application

Installation

1. pip install gesture_classifying_model
2. At top of pipeline.py insert line

   from gesture_classifying_model import GestureClassifier
   

3. Instantiate class in transfomers list, ex.:

   transformers = [
      AlreadyExistingTransfomer(),
      ...
      GestureClassifier(),
      ...
      AlreadyExistingTransfomer()
   ]
   

4. First run will cause download of weights to models/gesture_classifier with weights in *.h5 format and config.yml
5. Run appplication and start using system!

Classifier input info

• image size = [None, 64, 64, 3] (batch size, height, width, channels)
  • where None means, it can be arbitrary number of frames, only last frame is classified
• values range = [0, 255] in RGB (may be float or int)

Classifier ouput info

• only letters are classified
• available letters (both upper and lower)*:

abcdefghiklmnopqrstuwxy

• output size = [None, 24], where under None is same number as in the input

*the reason for such letters is that 'j' and 'z' both need movement, but classifier uses only singular frames

Requirements

Used environment:

• Python 3.5
• Theano 0.9.0
• Keras 1.2.2

Changelog

• v0.1.3:

  • documentation on PyPI

• v0.1.2:

  • first PyPI availability
  • downloading weights after one day, after next run