deep_hash_search

Image search based on convolutional neural network feature extraction.

Instructions

Download the dataset and extract it:

wget -c http://data.csail.mit.edu/places/places365/train_256_places365standard.tar
tar -xf train_256_places365standard.tar

Run to split the dataset into training and testing parts:

./train_test_split.sh

Training:

To train with hard mining (web api available):

 python3 resnet.py

After the model is trained and saved, you can run the sample website: FLASK_APP=web_pova.py flask run

To train without hard mining (web api not available):

  python3 model_noHardMin.py

To evaluate the saved trained model run:

python3 eval_results.py

Requirements: Keras, tensorflow, [Flask - for web api]

Authors:

• Tomáš Sýkora (tms.sykora@gmail.com)
• Josef Jon (xjonjo00@stud.fit.vutbr.cz)