
CASE STUDY
Fashter
Machine Learning Powered Visual Search for Fashion.
Transforming the online shopping experience with cutting-edge image recognition technology.
CHALLENGES & SOLUTIONS
About the project
Project Overview
Meet Fashter, an avant-garde clothing search platform powered by Machine Learning and image recognition technology. Fashter revolutionizes the way users shop online by offering visually similar clothing options based on user preference, thereby providing a personalized and simplified shopping experience.
Key Application Features:
Intuitive Search Bar: We developed an easy-to-use search bar where users can input their desired clothing items to generate initial results, making the search process direct and uncomplicated.
Inspiring Random Selection: For those unsure of what they want, we implemented a feature that presents a random selection of clothing items, offering inspiration and a potential starting point for their search.
Visual Similarity Search: Leveraging Machine Learning technology, we designed a ‘Similar’ feature. When a user selects an item they like, the application generates a collection of visually similar items, offering a wide range of choices that align with their personal style.
Streamlined Shopping Experience: Once users find their ideal clothing item, they can visit the retailer’s website by clicking the ‘Store’ button. This feature provides a seamless transition from searching to purchasing, enhancing the overall shopping experience.
Ad-free User Experience: In our pursuit to ensure an undisturbed and focused user experience, Fashter is designed without popups or ads. This decision underscores our commitment to prioritizing user needs and preferences.
The Journey of Building Fashter
The inception of Fashter began with an aim to enhance the online shopping experience by creating a visually-based clothing search platform. Our vision was a tool that not only matches users’ preferences but inspires them by providing similar fashion alternatives.
Our team selected Ruby on Rails as the development framework for its efficiency and ease of use. Machine Learning was incorporated to handle the visual recognition and similarity mapping tasks. The application was deployed on Heroku, given its reliable, scalable, and agile cloud platform services.
The ‘Similar’ feature posed a unique challenge – creating a Machine Learning model capable of accurately identifying and suggesting visually similar clothing items. Our solution was to train a model using vast datasets of clothing images, ensuring it could recognize patterns, colors, and styles with high precision. This advanced feature brings immense value to the tool, distinguishing Fashter from conventional online shopping platforms.
The Power of Fashter
Embarking on a Personalized Search: Fashter’s search bar or random selection features offer users a personalized and unique starting point. Whether they know what they’re looking for or need a little inspiration, Fashter caters to all shopping moods.
Discovering Visually Similar Options: When a user spots an appealing item, the ‘Similar’ feature displays a range of visually similar clothing pieces. This offers users an array of alternatives aligning with their taste, enhancing their browsing experience.
Seamless Transition to Purchase: After identifying their ideal item, users can seamlessly transition to the retailer’s website by clicking ‘Store.’ This streamlined process promotes an effortless shopping experience, from discovery to purchase.
Experience Ad-free Shopping: Fashter’s commitment to providing an undisturbed and focused user experience is evident in its ad-free design. This approach respects users’ time and focus, ensuring their interaction with the platform remains purely about fashion and personal style.
-
Software engineering
& development -
UX/UI design
& development -
Applications, website,
SaaS


OTHER SOLUTIONS
Case studies

University Impact Evaluation Tool
Machine Learning and Semantic Search powered tool for gathering, processing, and visualising data on the impact of research institutions across economic sectors.
View →


Statometras
A web app that allows people to see the most important changes of employee count in every Lithuanian company.
View →

University Impact Evaluation Tool
Machine Learning and Semantic Search powered tool for gathering, processing, and visualising data on the impact of research institutions across economic sectors.
View →