Web-Based Model for Apparel Classification

3 Pages Posted: 25 Apr 2019

Date Written: April 8, 2019

Abstract

Clothing classification has its application in e-commerce and advertising. With the advancement of fashion industry, it becomes necessary to introduce a new method for online shopping. This paper describes how a convolutional neural network (CNN) can be used to classify garments from a given dataset. The final aim of the project is to identify apparels which may or may not be worn by an individual and to give a list of similar apparels. For this purpose, a CNN model is trained on the relevant dataset with the help of tensorflow. Later, a web application is created using tensorflow.js wherein a user can upload an image and the model will identify the image using the labels in the dataset and display it on the webpage. Further, this web application will be converted into an android application which is optimized for smaller screens for the ease of accessibility.

Keywords: Convolutional Neural networks(CNN), Test Dataset, Validation and Training Dataset, Tensorflow, Node.js

Suggested Citation

Bhimani, Harsh and Kaimaparambil, Keerthy and Papan, Vishal and Chaurasia, Harsh and Kukreja, Amit, Web-Based Model for Apparel Classification (April 8, 2019). 2nd International Conference on Advances in Science & Technology (ICAST) 2019 on 8th, 9th April 2019 by K J Somaiya Institute of Engineering & Information Technology, Mumbai, India, Available at SSRN: https://ssrn.com/abstract=3367732 or http://dx.doi.org/10.2139/ssrn.3367732

Keerthy Kaimaparambil

Independent ( email )

Vishal Papan

Independent ( email )

Harsh Chaurasia

Independent ( email )

Amit Kukreja

Independent ( email )

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
108
Abstract Views
680
Rank
454,063
PlumX Metrics