Image-Retrieval-With-User-Feedback

About: with picture cut and the user feedback based on SVM, this is the final course design in my university, i do this project with a lot of other's open source code, show my thx to all these guy~

requirement

keras
python3
h5py
sklearn
You can install all by the command (pip3 install *** )

File structure

├── README.md
├── index.py
├── face_cluster
│   │ 
│   ├── controller.py
│   ├── controller.pyc
│   ├── controller_det.py
│   ├── controller_image.py
│   ├── controller_image.pyc
│   ├── extract_cnn_vgg16_keras.py
│   ├── feedback.py
│   ├── settings.py
│   ├── settings.pyc
│   ├── urls.py
│   ├── urls.pyc
│   ├── wsgi.py
│   └── wsgi.pyc
├── manage.py
├── static
│   ├── img
│   │   ├── 12_t.jpg
│   │   ├── 13_t.jpg
│   │   ├── 14_t.jpg
│   │   ├── 15_t.jpg
│   │   ├── 16_t.jpg
│   │   ├── dataset1----------------------------------------------this file contain your own image
│   │ 
│   │ 
│   ├── index file --------------------------------------this file is your index file,you can get it from the train step
│
│
└── template
    ├── feedback.html
    ├── home.html
    ├── home2.html
    ├── photocutter.html
    ├── search.html
    ├── search.html.det
    ├── search2.html
    └── search3.html

Usage

First. Train The index file

In this step you should train a index file for your own image database, you can get it by python3 index.py -database <path-to-dataset> -index <name-for-output-index> the is the path to your own dataset1. then it will generate a index file, and you should put it into the (static/) file. maybe in this step you should download some file like vgg16_weights_tf_dim_ordering_tf_kernels.h5 etc. please wait....

Run The System

In this step, you should can use python3 manager.py runserver and you can use visit 127.0.0.1to see the web.

The System demo picture

This is the main pic

This is the search pic

This is the result pic

This is the feedback_1 pic

This is the feedback_2 pic

This is the feedback_result pic

This is the pic_cut pic

Remember star🌟 me, thx~😊