Inspiration.
Wherever you go
Decoir will change the way you approach interior design.
See what it can do for you
UOWD Innovation Fair People’s Choice Award
1st
3rd
UOWD Innovation Fair Judge Panel Vote Award
Create personalized moodboards and
discover furniture we know
you’ll love
Decoir is a graduation project realized by a team consisting of five passionate computer science graduates Yuliya Volkova, Maryam Zia, Hadi Daou, Nawwaf Husain, and Calwyn Pereira, specializing in cybersecurity and mobile and game development.
Conceptualized in September 2022, Decoir intends to provide a fresh take on the interior design brainstorming process.
An elegant tool for your
interior design ideas
Seamless integration
Download on iOS or Android
and use conveniently with your
account to never lose
your ideas.
Identifies on the spot
Take a picture on a whim or upload from your gallery to identify the make of your
furniture efficiently.
Furniture you love
Over 8,700 pieces of furniture from popular retailers in UAE will guarantee we will find exactly what you need.
Simplifies purchases
In-built visual storage for your
saved furniture enables you to
directly purchase it from the retailers.
Automate moodboards
Select the furniture you’re inspired by and Decoir will generate personalized moodboards for your taste.
Take creative control
Use our extensive furniture catalog to edit or create your own moodboards for a perfect personalized interior.
Share with friends
Export moodboards in various formats to your friends and
clients with a simple tap of
a button.
Expedite your workload
Decoir provides an out of the box solution with tools targeted at interior designers looking to maximize their productivity.
Behind the scenes of
Decoir
Machine learning and deep learning models were employed to classify the input furniture images and to generate stylistically accurate mood boards. Two CNN models and a K-NN algorithm model were employed to classify the user’s input furniture image to the closest image match within Decoir’s furniture dataset.
To identify the user’s style from an input of three to five previously saved furniture images to generate mood boards, a trained K-means clustering model and a CNN model were employed to return stylistically similar images which will be used in the automatic generation of mood boards.
To procure Decoir’s dataset consisting of 8,700 furniture models, enterprise-level scraping applications as well as open-source python-based scraping frameworks were used interchangeably.
