dc.contributor.author |
Bandara, Uthpala |
|
dc.date.accessioned |
2024-03-12T06:58:58Z |
|
dc.date.available |
2024-03-12T06:58:58Z |
|
dc.date.issued |
2023 |
|
dc.identifier.citation |
Bandara, Uthpala (2023) A System for Outfit Recommendation . BSc. Dissertation, Informatics Institute of Technology |
en_US |
dc.identifier.issn |
20191305 |
|
dc.identifier.uri |
http://dlib.iit.ac.lk/xmlui/handle/123456789/1840 |
|
dc.description.abstract |
"In the fashion business, personalized outfit suggestions have grown in popularity as shops explore for new methods to give customers an engaging and personalized buying experience. The lack of readily available pre-designed clothes that suit specific consumers' demands and tastes is one of the biggest problems with outfit recommendation systems. paragraph two
Author created an outfit recommendation system using ResNet50, a well-known neural network architecture for image classification, and Generative Adversarial Networks (GANs) to solve this problem. The ResNet50 model is utilized in this system to categorize the created outfits and suggest the most appropriate ones to the user. This system employs the GAN to generate new outfit combinations that correspond to a user's preferences.
Author carried out a user study where a group of volunteers were asked to score the appropriateness and general attractiveness of the suggested outfits in order to assess the effectiveness of our outfit recommendation system. Authors findings demonstrate that, in terms of accuracy and user happiness, Authors approach performs better than a number of cutting-edge outfit recommendation systems.
" |
en_US |
dc.language.iso |
en |
en_US |
dc.publisher |
IIT |
en_US |
dc.subject |
Outfit recommendation |
en_US |
dc.subject |
Generative Adversarial Networks |
en_US |
dc.subject |
User satisfaction |
en_US |
dc.title |
A System for Outfit Recommendation |
en_US |
dc.type |
Thesis |
en_US |