Wardrobe

Wardrobe: Your personal AI-Stylist Dress for the weather, dress for the occasion, dress to impress.

Comment

Inspiration

Have you ever felt under or over dressed at an event? Or caught a cold because you didn’t wear enough? Maybe you don’t know what to wear today?

Well we have a personal stylist for you accessible anytime, anywhere. “Wardrobe” revolutionises the way we approach clothing. It’s a smart tool that integrates AI to help you pick the perfect outfit for any occasion, weather and mood. Plus, it makes fashion fun and social. Share your style with friends, plan outfits together or seek inspiration from people all over the world.

We created “Wardrobe” to eliminate the time and worries of picking the right outfit, making every day a perfect outfit day for everyone.

What it does

Wardrobe is an interactive web app that transforms your clothing experience. Simply upload pictures of your clothes, and let the AI stylist work its magic, offering personalised recommendations on fits elevating your fashion game. Wardrobe is designed to simplify your daily outfit decisions. Whether you’re dressing for a specific occasion, looking to refresh your daily outfits or layering up for the cold weather, let Wardrobe do the hard work of

  • Features
    • Keep Track of Your Wardrobe:
      • Every time you upload a picture, store that piece of clothing in your own virtual wardrobe.
      • Clothes in the wardrobe can be used in future fit suggestions
    • Browse past fits:
      • Favourite and keep track of past fits
      • Categorise them based on tags, season, occasion and mood.
    • See what others around the world are wearing or your friends
      • Upload and share fits so your friends can see
      • Or find other people around the world with similar fashion tastes.
    • Suggestions:
      • When you upload a piece of clothing, the AI can suggest similar pieces of clothing or pieces of clothing that would pair well.
    • This feature is personalised to the user.
      • The AI will learn from your virtual wardrobe, past liked fits, searches.

How we built it

Building "Wardrobe" involved the use of technical tools, design and planning. It began by identifying how we could utilise emerging technologies to improve and streamline day-to-day life. Thus, the concept of Wardrobe came to fruition. Picking and trying on outfits can be a time-consuming and tiring process, we wanted to provide a digital solution that could revolutionise the way people interact with their wardrobes. Given the theme “Shaping the Future”, we couldn’t resist using AI. By blending AI with fashion, our aim was to offer personalised styling recommendations that could cater to individual tastes, occasions and weather conditions.

During the design phase, we focused on creating an intuitive and engaging user experience; prototyping on Figma for visualising and testing. Just like how we wanted to simplify the process of picking an outfit, our website design and colour palette was kept minimal and monochrome so users could focus on the clothing and the fit itself. Which is why the UI was crafted to be minimalistic yet functional allowing users to easily upload photos, browse their virtual wardrobe and interact with outfit suggestions.

In order to deliver a seamless and intuitive user experience, we harnessed a robust and modern technology stack:

  • Frontend: ReactJS for the web application, TailwindCSS and AntD for styling, hosting on Firebase Deploy

  • AI/Machine Learning: PyTorch and K-means.

  • Backend, Deployment and Database: Google Firebase in Python

At the heart of our project lies our AI model. By combining multiple Machine Learning and Deep Learning Algorithms, we were able to create an AI model that uses the deep learning python package “Pytorch”.

Pytorch leverages a prebuilt image backgroun deep learning algorithm to enhance image processing tasks. It utilises a K-Means clustering model to analyze the colors of clothing items, identifying the most prominent colors with precision. Furthermore, PyTorch employs an advanced deep learning neural network framework to accurately identify different types of clothing, showcasing its capability to handle complex image recognition tasks.

Our commitment to innovation, combined with a responsive approach to user needs, ensures "Wardrobe" remains at the forefront of digital fashion solutions, making everyday styling easier and more enjoyable for everyone.

Our Sustainable Business Model

  • Affiliate Marketing: Earn commissions through partnerships with sustainable fashion brands when users make purchases based on app recommendations.
  • Sustainability Certifications: Offer educational courses on sustainable fashion practices, awarding certifications upon completion.
  • Retail Partnerships and Collaborations: Collaborate with influencers and designers for exclusive content, splitting profits or charging sponsorship fees.
  • We have chosen this business model to ensure “Wardrobe" thrives as a free web app while still adhering to a sustainable practices.

Competitors and how to overcome

Competitors in the Market:

  • Stick Fix: Offers personalised clothing subscriptions using a mix of stylists and AI
  • Cladwell and Stylebook: Provide outfit planning and wardrobe management but with limited AI personalisation.

What sets us apart from competitors:

  • "Wardrobe" combines AI-driven styling with a focus on sustainability, offering personalized outfit recommendations based on the user's existing wardrobe. It allows users to take a picture or upload an existing picture anytime and anywhere.
  • "Wardrobe" encourages sustainable fashion practices, aligning with growing consumer awareness and preference for eco-friendly options.

Challenges of the product in the market and how we will address these:

  • Ensuring user data privacy and security - Implement encryption and transparent privacy policies to build trust.
  • Maintaining high engagement and daily usage - Introduce interactive community features, chatting, following and followers, and allow users to personalise their own wardrobe board to boost daily engagement.
  • Differentiating in a crowded market of fashion tech apps - Leverage the unique selling point of promoting sustainable fashion choices and personalised styling advice based on existing wardrobe items, setting "Wardrobe" apart from competitors focusing on new purchases.

Challenges we ran into

During the development of “Wardrobe" there were many situations and challenges that pushed us to innovate and adapt. One of the initial hurdles was making sure the AI knew what was “fashionable” and what was not, it needed to feel intuitive and seamless for users. We needed extensive datasets, iterations and GPU power to train the AI model in order to accurately analyse and suggest outfits. This took over 30 hours and more to train the model ensuring recommendations were not just relevant but also personalised and diverse.

Another challenge we faced was in the UI/UX. Creating a user-friendly interface and being able to style that in code was something we frequently discussed within the group. On the technical side, the code had to handle the complexity of user interactions such as navigation, search and uploads while also ensuring the web app’s performance remained swift and responsive, especially when processing large volumes of image data and files.

In terms of UI, it was pivotal to balance functionality with simplicity in our design minimising distractions and delivering the information needed to the user without them being overwhelmed.

Drawing inspiration from successful and well-designed websites such as apple.com , nike.com and endclothing.com led to creative solutions. We were able to leverage user-centric design patterns while ensuring the navigational experience was still rich and intuitive.

By continuous cycles of iteration, testing, and feedback, incorporating insights from user behavior to refine the application, we were able to make rapid adjustments and feature enhancements in response to real-world user interactions.

Accomplishments that we're proud of

AI Integration: Successfully training and integrating an AI stylist offering personalised fashion advice. The ability to analyze user-uploaded images and suggest outfits that cater to individual styles, occasions, and weather conditions showcases our advanced use of machine learning and AI in practical, user-centric applications. Not only that, but to be able to train and fully integrate an AI model into a working web application in under 48 hours has made us incredibly proud.

  • The accuracy of the machine is over 80% as well, making it on track to become a genuinely useful and reliable tool.

A Working Full-Stack Web Application: To successfully launch a fully operational full-stack web application, integrating ReactJS and TailwindCSS for the frontend with Firebase handling the backend and database operations is something we take pride in. This accomplishment not only demonstrates our ability to create a comprehensive and cohesive system but also showcases the teamwork and dedication needed to bring “Wardrobe" to life.

UI/UX: Our UI delivers an aesthetically appealing and functionally intuitive user interface. The monochrome theme and large white spaces contrasting clothing pieces focuses the users attention. We have created a minimalistic yet compelling interface that simplifies fashion decisions for a global user base. The website feels calm and not too overwhelming, we wanted the user to only think about clothing and not worry about navigation and account logins. The use of the parallax landing page allows for seamless transitions into additional features of the website such as the explore page without the user having to search for it. The idea was to provide an immersive journey, guiding the user around the webpage.

Our Team: To be able to present a fully functional product, collaborate, communicate and chat (or maybe friendly disagreements sometimes) and still come out as a stronger team makes us incredibly proud.

What we learned

Throughout the development of “Wardrobe", we gained invaluable insights across multiple facets.

From the technical side, delving into AI and machine learning enhanced our skills in big data and cloud computing. Being able to integrate an AI into a full responsive, user-centric web application was something that none of us had done before, making it a challenging but fulfilling learning experience.

The design and user interface was also a challenge that we tackled. Our team poured lots of effort into learning essential design principles and analysing aesthetically pleasing websites with good UI/UX. We found that simplicity and functionality are key to engaging users effectively. This insight was pivotal not just for user satisfaction but also for the longevity and scalability of our app.

Having adopted an agile development process, we also learned the significance of flexibility and rapid iteration based on testing and feedback within our team. Training the model for better accuracy and having our team testing the web app was essential to creating a viable product.

From a more personal side, we rediscovered our passion for building a impactful and enduring app, this hackathon has brought out our teams resilience and perseverance in order to overcome challenges.

What's next for Wardrobe

“Wardrobe” Roadmap towards sustainable growth and impact:

  • Short-term Objectives (1-6 months)
    • AI and UX improvements: Refining AI algorithms for more personalised recommendations. Optimise the UI/UX
      • Prototype testing, adjusting features, improving the AI’s accuracy and enhancing user engagement
    • Focus on security and user privacy such as data encryption, secure authenticaion process when logging in and regular security audits to protect our users.
    • Gather community engagement and feedback through social media, forums and early access programs to gather insights and foster a sense of community.
    • The feedback loop is instrumental in improving and updating the web app.
  • Mid-term Objectives (6-12 months)
    • Partnering with fashion brands and influencers to promote sustainability. And keep developing features that support sustainable practices, like having our AI find and recommend clothes from thrifting websites such as facebook market place or depop.
    • Hire professional stylists or trending fashion influencers to help train the AI and offer a more human touch to our product.
  • Long-term Objectives
    • Grow our user base and community. And have our web app be able to scale for a larger number of users.
    • Keep strengthening our security measures and data privacy.
    • Continue to add in new features promoting sustainable growth and enhancing our user experience.

Future features: These features are aimed to enhance user engagement and experience, promote sustainable fashion and circular fashion.

  • Trending pieces of clothing page
    • Allows users to see what is trending based on their location.
    • Can also filter based on clothing types, styles and colour.
  • Augmented Reality Try-on: Allow users to scan themselves to upload onto the website and have the AI generate images with the user wearing the clothes they select.
  • Sustainable Impact Tracker: Feature that displays the environmental impact of the user’s wardrobe choices, encouraging sustainable fashion choices.
  • Wardrobe Swap: Feature where users can find others to swap their clothes. Extending the life of garments
  • Community Fashion Shows: Users can organise or participate in virtual fashion shows within the app fostering community engagement.
  • Premium Subscription: Offer additional features like advanced AI styling, expert consultations that include personalised sessions with professional stylists for outfit planning and wardrobe management. And an ad-free experience all for a monthly fee.
Share this project:

Updates