Clothero

• Introduction

  <ul dir="auto">
    <li>
      AI-based shopping assistance for both man and woman using transfer learning and statistics like Cosine and Pearson similarity metrics. The basic idea behind the project is passing image data through a Pre-trained model [Mobilenet_v2] for extracting feature vector which is flattened out in the end. The output from this goes through a comparison metric called cosine similarity metric which is computed between users choice of cloth and the database vectors, scrapped from amazon.in
    </li>
  </ul>
  

• Pre-requisites

  <ol dir="auto">
    <li>
      Transfer Learning<br /> here we have used pre-trained model for feature extractor ,particularly
    </li>
    <li>
      Statistics<br /> Basic knowledge about similarity metrics like Cosine , Pearson
    </li>
    <li>
      Web-scrapping<br /> using webscraping various shopping sites for data for building the prototype using scrapy
    </li>
    <li>
      Web-technologies<br /> Baiscs of CSS, HTML, JAVASCRIPT , Flask ,JQuery
    </li>
    <li>
      Database<br /> No-sql database like Mongodb
    </li>
  </ol>
  

• Installation requirements

  <ol dir="auto">
    <li>
      Python3 <ul dir="auto">
        <li>
          numpy
        </li>
        <li>
          tensorflow
        </li>
        <li>
          pandas
        </li>
        <li>
          pymongo
        </li>
        <li>
          scrapy
        </li>
        <li>
          math
        </li>
      </ul>
    </li>
    
    <li>
      Databse <ul dir="auto">
        <li>
          Mongodb
        </li>
      </ul>
    </li>
    
    <li>
      Backend <ul dir="auto">
        <li>
          Flask
        </li>
      </ul>
    </li>
    
    <li>
      Front <ul dir="auto">
        <li>
          CSS
        </li>
        <li>
          HTML
        </li>
        <li>
          JAVASCRIPT
        </li>
        <li>
          JQuery
        </li>
      </ul>
    </li>
  </ol>
  

• Intallation

  <ol dir="auto">
    <li>
      Clone the repository
    </li>
    <li>
      Install mongodb
    </li>
    <li>
      Preprocess data i.e create vectors of images path with the help of <code>Clothero.ipynb</code> in the root directory and sample data csv for preprocessing can be found in <code>img-databse</code> folder.
    </li>
    <li>
      Import processed csv file in your mongodb database
    </li>
    <li>
      Install the requirements</p> <ul dir="auto">
        <li>
          for Windows users</p> <blockquote>
            <p>
              pip install -r requirements.txt
            </p>
          </blockquote>
        </li>
        
        <li>
          for Ubuntu users</p> <blockquote>
            <p>
              pip3 install -r requirements.txt
            </p>
          </blockquote>
        </li>
      </ul>
    </li>
    
    <li>
      Start the server<br /> <code>cd server</code><br /> <code>flask run</code>
    </li>
  </ol>
  

• License

• Reach us

• Ram Suthar
  LinkedIn | Github
• Krish Shah
  LinkedIn | Github