Aesthify

The Project

For my project on the photo-taking assisting design, it was initially for another purpose. Since I have always been an enthusiast in shooting photos and learning photography, I have had moments where I thought the photography learning experience could be improved drastically. Since the art of photography is subject and subtle like every other media of arts, sometimes I could not comprehend specifically what makes a photo good, especially when it is from one of the photography masters. This hinders the development for a better sense of aesthetic for photographers. More often, after taking the shots, it is even harder for photographers themselves to determine whether if it is genuinely good. Without constant feedback, gaining improvements in photography is difficult and uncertain, unless people have a teacher or a professional photographer to go to at any time. Hence, I wanted to explore the possibility of quantifying aesthetic and letting AI to score photographs for feedback in a mobile app. This idea of improving the photography-learning experience was my initial design. After this, I began the journey of studying machine learning and developing the app with a friend of mine with prior experience in app development. First, I designed the UI for the app and studied how to implement the UI through a coding framework called Flutter. Then we managed to install a model adapted from a Kaggle notebook and trained it using the AADB dataset on a platform provided by PyTorch. Finally, after overcoming numerous issues like converting a machine learning library to another, granting permission for cameras on Iphone, and halting when loading photos, we succeeded in getting a basic prototype running. The app can take photos using its camera feature or upload a photo to receive detailed feedback on multiple aesthetic attributes like lighting, color harmony, balance, and more. This feedback highlights specific areas for improvement, empowering users to easily reflect on their photos and make adjustments for shooting again. Additionally, the app includes a photo gallery where users can see their photos reviewed. The app so far allows photographers and photography learners to conveniently develop their photography aesthetics and skills through reviewing feedback on photos in anytime. Nevertheless, an idea had come to my mind while developing the app. I thought if feedback of photos can help people improve on photography skills, it can also help people take better photos on the spot. In today’s world where social media is striving, countless people have the demand for documenting their lives with better photos, and they don’t necessarily have to be enthusiasts for photography. That's exactly why, with the exisiting function of this project, I would like to serve it as an asisstant for taking better photos, in addition to a tool for developing photography skills.

Mobile

About the team

  • China

Team members

  • Trevor