Clothes detector, predicting patterns and color together with the clothing category. Made for my internship Summer 2018 @Somera.
Exemple Detections
Code uses Keras API Tensorflow as backend.
-
In order to run the code Tensorflow Object Detection API set-up needed.
-
Copy and paste all the files in to the tensorflow/models/reserach/object_detection directory
-
Dataset taken from: http://streetstyle.cs.cornell.edu/
-
clothes_detector.py (uses tenorflow object-detection API and reads images from static/img folder outputs the prediction in train_images folder)
-
train_test_splitter.py (clean, visualize the dataset dataset ready to feed into Keras classifiers)
-
prediction_module.py (runs custom classifiers on Tensorflow Object detection API's region proposals)
-
keras_models.py (creates color, pattern, clothing classifiers uses Keras API)