/Visual-Search-TA

Primary LanguageJupyter Notebook

Hello Buddy😎

Pencarian Hats Hasil Proses Pencarian Hats

Table of Contents

  1. Introduction
  2. Dataset
  3. Implementation
  4. How to launch
  5. Note

Introduction

This repository is the result of the implementation of my Final Project (Thesis) at the Yogyakarta Technological University in pursuing a bachelor's program in the department of Informatics. The title of my thesis is "IMPLEMENTATION OF DEEP LEARNING FOR PRODUCT IMAGE SEARCH SYSTEMS USING CONVOLUTIONAL NEURAL NETWORK (CNN) ALGORITHM".

Dataset

The dataset used is the Real Fashion dataset from the kaggle website. The Real Fashion dataset contains 30,276 image data and consists of 7 categories with a total of 42 sub-categories in it. The 42 sub-categories in the dataset consist of objects commonly found in the marketplace in the fashion category.

Sampul Dataset

Informasi Dataset :

Type Information
Source Kaggle Dataset : Real Fashion
License Unknown
Category clothing and accessories
Rating 4.4
File Type and Size Folder (2 GB)

Implementation

Installation

  • Clone this repo
  • Install Python dependencies
$ git clone https://github.com/YogiDwiAndrian/Visual-Search-TA.git
$ cd Visual-Search-TA
$ pip install -r requirements.txt

Materials

Models

The model is used to classify categories in the search process and is also used for image feature extraction which will then be added to the database. The model can be obtained from downloading the model that I have trained (Model Resnet) or you can train yourself by running the Tugas_Akhir.ipynb program on Google Collaboratory. put the model in static/model/your_model.h5

Visual-Search-TA/
                |->static/
                |        |->data/
                |        |->dataset/
                |        |->model/
                |        |       |->model_resnet.h5
                |->templates/
                |            |->index.html
                |->connection.py
                |->data2dataset.py
                |->...

Datasets

  1. Download Kaggle Dataset : Real Fashion and extract it into the static/data/ folder
Visual-Search-TA/
                |->static/
                |        |->data/
                |        |       |->(Extracted data)
                |        |->dataset/
                |        |->model/
                |->templates/
                |->connection.py
                |->data2dataset.py
                |->...
  1. After that, changing the structure of the data that originally contained a main-category, then each main-category containing sub-categories was changed to only sub-categories. Run the script data2dataset.py
$ python data2dataset.py
Visual-Search-TA/
                |->static/
                |         |->data/
                |         |      |->dress/
                |         |      |        |->mini
                |         |      |        |->midi
                |         |      |        |->...
                |         |      |->...
                |         |->dataset/
                |         |       |->mini
                |         |       |->midi
                |         |       |->...
                |         |->model/
                |->templates/
                |->connection.py
                |->data2dataset.py
                |->...
  1. Adding feature extraction data and images to the database. Run the script image2db.py
$ python image2db.py

OR

You can skip steps 1, 2, and 3 by downloading the sql data from the export data that I did ta.sql. After successfully downloading it then just import it into your database.

How to launch

The program is run using the Flask web framework, to run it by running the main.py script.

$ python main.py

Then we just open it in our favorite browser, http://localhost:5000/ or http://localhost:5000/index, it will display the program page that is run. To stop the running Flask program, just type Ctrl+C at the terminal/command prompt

Note

📞📬 Stay in touch

LinkedIn Instagram