Smart Closet: AI integration in smart home appliances

You’ve probably experienced indecision while picking out your outfit… once at least!

So, you might relate to the innovators’ question of: “Why can’t we all have our own stylist at home?”

With technology, it’s possible, making the experience more fun and far easier.

The Smart Closet is a new way of picking your outfits. It simplifies the user’s experience by only requesting a voice command on the type of outfit needed and providing it in a few seconds. Say, for example, a formal outfit for a summer occasion, or a workout outfit for a spring day, it will be handed to you in moments!

As part of the Young Innovators Lab (a program launched in collaboration between Thinking Oasis and the Abdul Hameed Shoman Foundation), the Smart Closet project stands as a testament to what young minds can achieve when given the right tools and guidance. The yearlong training journey was designed, developed, and conducted by Thinking Oasis; focusing on nurturing innovation, critical thinking, and real-world problem solving.

This smart closet project was conceived, innovated, and implemented from A to Z—from the initial idea, through proof of concept, to a fully functional prototype—by one group of passionate participants in the lab, all aged between 10 and 13 years old.

Meet the young innovators behind the Smart Closet project:

The closet has built-in AI-powered software that is trained on thousands of photographed clothing items and outfits, gaining the ability to recognize the features of a new item using just a camera image. It might label a new piece as “a black formal winter coat,” for example. It can then put together an outfit from items that fit well together and match the occasion. A microphone in the closet hears the user’s voice commands and analyzes them, then activates the internal mechanism to hand over the right items.

‍

The Problem beyond fashion choices: helping the visually impaired

What started as a frustration with daily indecision evolved into a much deeper question: How can technology support people in choosing and coordinating outfits—especially those who can’t see them? For visually impaired individuals, picking clothing independently is often a struggle. This project tackles challenges head-on.

‍

Audience & Impact

The Smart Closet serves a broad audience—from everyday users wanting an easier morning routine to visually impaired individuals who need accessible solutions. The project doesn’t just help pick clothes; it empowers independence, restores confidence, and makes style inclusive. It’s a fusion of AI, accessibility, and daily convenience.

‍

The Concept’s Origin

The concept was born from a mix of empathy and curiosity. The young innovators imagined how smart technology could ease decision-making and expand autonomy. They wanted to move beyond traditional smart mirrors and virtual fitting rooms towards an actual physical system that delivers results, not just suggestions.

‍

How They Built It

The young innovators used a supervised machine learning model to classify clothing types and determine their suitability for various occasions. They started by preparing and refining a custom dataset, labeling images to help the model recognize patterns and learn the visual characteristics of different clothing items. Through iterative training and testing, they optimized the model's performance and successfully achieved a precision of 98.3% and a recall of 98.4%, in recognizing and classifying clothing from photos.

Each clothing item added to the closet is photographed by an internal camera. The AI labels it, stores its location, and remembers it. And that makes the closet knows the items on it all the time.

When the user interacts with the system (currently using buttons), the AI scans the inventory and selects appropriate items based on the request. Then comes the magic: the mechanical system brings the items to the user.

‍

Technical Breakdown

- AI Model: Trained on a custom image dataset to detect type (e.g., T-shirt, coat) and formality level (casual, formal, etc.).

- Camera & Classification: A camera captures a new item, which the AI classifies and locates.

- Hardware System:
A circular rotating rack holds the items.
Each hook includes a limit switch to detect whether it’s occupied.
A servo motor at the center rotates the rack until the right outfit is aligned for the user.

- User Input: Currently via buttons, with plans to include voice command processing for greater accessibility.

‍

Teamwork & Iterative Design:

The team’s journey wasn’t linear. They explored various machine learning platforms and tested multiple mechanical designs before arriving at the final structure. They collaborated through trial and error, improving their model iteratively—from dataset enhancement to software selection.

The mechanism was redesigned several times for practicality and functionality. At first, the moving system was too complex. After rounds of simplification and testing, they developed a feasible design using commonly available components.

Looking Ahead

Currently, the Smart Closet operates using buttons for user input, but the team aspires to reintroduce the voice recognition module, making it even more beneficial for the visually impaired. With voice commands, users wouldn’t need to see or touch the closet at all, making it fully accessible and truly intelligent.

Conclusion
The Smart Closet proves that artificial intelligence can be more than a trend—it can be a tool for inclusion, independence, and innovation. Whether you’re choosing what to wear or helping someone regain their independence, this AI-powered solution is a step forward for smart homes and smart humanity.

‍