IKEA furniture finder

Tan, Meng Sheng (2023) IKEA furniture finder. Final Year Project, UTAR.

Preview

PDF Download (10Mb) | Preview

Abstract

The aim of this project was to develop an app that provides users with furniture recommendations based on the category and color of the furniture they are interested in. Deep learning techniques were used to train a TensorFlow model to accurately classify images of furniture. The model was trained on a large dataset of furniture images that were collected and labeled using a PowerShell script for automatic dataset labeling. A Flask web application was built using this model to predict the category of furniture in images sent by the Android client app. Additionally, a REST API endpoint was implemented in Flask to retrieve random furniture images from Firebase, which were used to display recommendations to users. To ensure scalability and consistency across environments, the Flask app was deployed on a cloud platform using Docker. User testing was conducted to evaluate the accuracy and usability of the app, and feedback was solicited from users to identify areas for improvement. Overall, the results demonstrate that the Ikea Furniture Finder app is an effective tool for assisting users in finding furniture based on their preferences.

Actions (login required)