roboticslaburjc / 2016-tfg-david-pascual

2016-tfg-david-pascual

Project Name: Deep Learning on RGBD sensors

Author: David Pascual Hernández [[email protected]]

Academic Year: 2016/2017

Degree: Degree in Audiovisual Systems and Multimedia Engineering

Mediawiki: http://jderobot.org/Dpascual-tfg

Tags: Deep Learning, Keras

State: Developing

Usage

digitclassifier is a JdeRobot component which captures live video and classifies digits in the sequence with a convolutional neural network built with Keras. In order to launch it with Python 2.7 you must install:

• JdeRobot (installation guide)
• OpenCV 3 (it will be automatically installed with JdeRobot)
• Keras (installation guide)

Links for downloading augmented MNIST datasets for training and Keras model for testing:

• HDF5 datasets
• LMDB datasets
• Keras model (4 conv. layers, early stopping w/ patience=5)

If you want to launch digitclassifier, open a terminal and run:

cameraserver cameraserver.cfg

This command will start cameraserver driver which will serve video from the webcam. In another terminal, access the location where you've cloned the repo and run:

python digitclassifier.py digitclassifier.cfg

That command should launch the component and you should see something like this:

More info: [http://jderobot.org/Dpascual-tfg]