ImageClassifier_TensorFlow-Improved-
Image Style Classifier using deep learning
This is an improved CNN for classifing image asthetic styles using tensorflow
Summarizing, this code does:
1 - gets images on folder and append it in a big array. image source: flicker groups and pinterest. Download dataset used at: https://drive.google.com/drive/folders/0BznDcpqi4yXxbEc0NkdZVzdaUGc?usp=sharing
2 - convert images in one big tensor of color values
3 - set cnn parameters
4 - define network model
5 - set a array for labeling and compare each image style first version: the styles are noir, bright and long-exposure.
5 - set training
6 - run the function
This code was developed under the guideness of the college professor Lucas Lattari (github profile: lucaslattari)