Skip to content

Latest commit

 

History

History
64 lines (52 loc) · 2.32 KB

README.md

File metadata and controls

64 lines (52 loc) · 2.32 KB

Clothero

+

  • Introduction

    • 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
  • Pre-requisites

    1. Transfer Learning here we have used pre-trained model for feature extractor ,particularly
    2. Statistics Basic knowledge about similarity metrics like Cosine , Pearson
    3. Web-scrapping using webscraping various shopping sites for data for building the prototype using scrapy
    4. Web-technologies Baiscs of CSS, HTML, JAVASCRIPT , Flask ,JQuery
    5. Database No-sql database like Mongodb
  • Installation requirements

    1. Python3
      • numpy
      • tensorflow
      • pandas
      • pymongo
      • scrapy
      • math
    2. Databse
      • Mongodb
    3. Backend
      • Flask
    4. Front
      • CSS
      • HTML
      • JAVASCRIPT
      • JQuery
  • Intallation

    1. Clone the repository

    2. Install mongodb

    3. Preprocess data i.e create vectors of images path with the help of Clothero.ipynb in the root directory and sample data csv for preprocessing can be found in img-databse folder.

    4. Import processed csv file in your mongodb database

    5. Install the requirements

      • for Windows users

        pip install -r requirements.txt

      • for Ubuntu users

        pip3 install -r requirements.txt

    6. Start the server cd server flask run

  • License

License: MIT