Personal Stylist Test phase, srs, design phase and source code final deliverable
Project Domain
Deep Learning/ Image Processing
Introduction
The Personal Stylist is a project that aims to develop a styling assistant that can recommend personalized fashion choices to users. The system will use advanced technologies such as computer vision, machine learning algorithms, and natural language processing to understand users’ preferences, body type, and fashion styles. By using this system, users can improve their fashion choices, enhance their confidence, and make more informed purchasing decisions.
Functional Requirements:
- The system must have an image recognition module that can identify clothing items and accessories from user photos. The module should use computer vision algorithms to detect features such as color, pattern, and style.
- The system must be personalized, allowing users to input their body type, measurements, and fashion preferences. The system should also use machine learning algorithms to learn from user feedback and adjust recommendations accordingly.
- The system must provide fashion recommendations to users based on their preferences and body type. The recommendations should include clothing items, accessories, and complete outfits.
- The system must integrate with social media platforms, allowing users to share their fashion choices and receive feedback from friends and family.
- The system should integrate with e-commerce platforms, allowing users to purchase recommended items directly through the application.
- The system must be scalable, allowing for easy integration with new fashion brands and expansion to support additional users and features.
- Make an interface on mobile application (android) from where a user can capture a picture and your trained model can recommend the style to the user.
Dataset
- DeepFashion Dataset:
This dataset contains over 800,000 images of clothing items categorized into 50 classes such as dresses, pants, shoes, and bags. It also includes attribute annotations such as color, texture, and style.
Link: http://mmlab.ie.cuhk.edu.hk/projects/DeepFashion.html
- Fashion MNIST Dataset:
This dataset contains over 70,000 images of 10 different types of clothing items such as t-shirts, dresses, and sneakers. It’s a popular benchmark dataset for evaluating machine learning algorithms related to fashion.
Link: https://github.com/zalandoresearch/fashion-mnist
- iMaterialist Fashion Dataset:
This dataset contains over 200,000 images of fashion products with attribute annotations such as color, pattern, and style. It also includes segmentation masks to identify the different parts of the clothing item.
Link: https://www.kaggle.com/c/imaterialist-fashion-2020-fgvc7/overview
- ModaNet Dataset:
This dataset contains over 55,000 images of fashion products with attribute annotations such as category, style, and occasion. It also includes segmentation masks to identify the different parts of the clothing item.
Link: https://github.com/eBay/modanet
Tools:
Android Studio, Python, Anaconda, OpenCV, TenserFlow, Keras.
Supervisor:
Name: Syed Aun Ali Bukhari
