Inspiration
https://www.pewresearch.org/social-trends/?attachment_id=34762 Pew Research Center lists barbers as among the groups of workers whose day-to-day job has the least exposure to AI.
To solve this problem, we propose an AI-powered mobile app which would allow users to preview potential haircuts together with a barber from the palm of their hands.
Background
We searched arxiv for papers involving hair strand reconstruction, pose alignment, and haircut transfers. We read and drew inspiration from the following papers: Neural Haircut: Prior-Guided Strand-Based Hair Reconstruction Style Your Hair: Latent Optimization for Pose-Invariant Hairstyle Transfer via Local-Style-Aware Hair Alignment Barbershop: GAN-based Image Compositing using Segmentation Masks
Ultimately, we decided to work with the code from the "Barbershop" and "Style Your Hair" paper to develop our backend.
What is kapper?
kapper is a cutting-edge solution which utilizes generative artificial intelligence to create virtual haircuts for users. Users may simply upload a photo of themselves and upload another image of their desired haircut style. The models developed by the authors of "Barbershop" and "Style Your Hair" then process the images, identifying the hair, and user poses, and transplants the desired hairstyle onto the user's photo, creating a realistic and personalized virtual haircut.
How we built it
kapper was developed using Python Flask as a backend, and React Native for frontend. kapper's backend is hosted on Google Cloud Platform. kapper's GAN models are written in Pytorch. We used React Native to create a user-friendly mobile app which allows users to easily upload their photos, select styles, and view their virtual haircuts.
Challenges we ran into
One of the major challenges we encountered during the development of KapperAI was deploying the machine learning model on Google Cloud.
Accomplishments that we're proud of
We are incredibly proud of several accomplishments in this project. Firstly, we successfully developed a full-stack application that empowers users, especially barbers, with the power of a GAN within the palms of their hands. We hope that in the future,
What we learned
Throughout the development of KapperAI, our team gained valuable insights into various aspects of app development and machine learning integration. We deepened our understanding of React Native, which allowed us to create an intuitive user interface. Moreover, we learned how to utilize various components of a complex application, including backend machine learning algorithms and frontend user experiences.
What's next for KapperAI
Our vision for kapper is to scale and expand its features. In the long future, we plan to incorporate more advanced machine learning techniques to offer even more realistic virtual haircuts. We also aim to collaborate with professional hairstylists to provide expert recommendations and styling tips to our users. Ultimately, our goal is for kapper to become a the goto tool for barber clients to find their next haircut, and to interact with their barbers.
Log in or sign up for Devpost to join the conversation.