Image Classification using Convolutional Neural Nets and Keras. This classifier uses LeNet-5, a pioneering 7-level convolutional network by LeCun et al in 1998 that classifies digits.
Requirements: Python 3.5 ,Keras 2.0.2 , Tensorflow 1.2.1 , OpenCV 3.3, numpy 1.11.0
$ python train_network.py --dataset images --model Bharath_Kumar.model
python test_network.py --model Bharath_Kumar.model --image images/examples/bk19.jpg
Add your training images in images/<Your Label>/
To train the model.
To train the model.
Model generated during thhe training.
Contains Images for Training. Use your owm data set here.
LeNet Architecture.
Trained using Backpropogation algorithm with stochastic gradint descent. This is a Binary Classifier.
-Train acc: 96.4665%
-Test acc : 88.5039%