CalcuHandsome

Men want to look good, but we often don't know where to start. Fashion can be daunting, but now it's a picture away

Comment

Inspiration

Men want to look good, but many of them don't feel like they have the time to look better or are even scared of taking style risks (we all know that one guy who has been rocking the same haircut since 2004). Many men stray away from fashion out of a lack of motivation or sometimes even worse - embarrassment. Those who do decide to dive into fashion find that there are just too many options and that sometimes it can be such a hassle just to find a shirt that they like or to pick a new hairstyle to commit to.

What it does

Our product takes the stress and time commitment out of staying fashionable. With CalcuHandsome, all you need is your iPhone camera to start looking better. You just take a photo of yourself, and our app will determine the hairstyle that looks best on you.

How we built it

We started with a user interface. We decided that a mobile app is the most convenient platform for this service. Our app runs your photo against our AI model we created using Microsoft Azure to determine your face shape. This can be a nuanced and tedious process to do by hand, so AI was the clear solution to this. From there, we provide the user hairstyles that work best with that face shape - guaranteeing that we suggest a hairstyle that is mathematically backed to make you look better.

Challenges we ran into

Our biggest challenges came from making the app UI and making the model accurate. We were all new to iOS development and the Swift programming language can be very stubborn when you don't follow Apple's rules exactly. Some pages took hours when it could have taken minutes on a web page; however, this was worth it in order to make the whole experience much more enjoyable and for us to learn how to develop iOS apps.

Making the model also proved to be very difficult. Many models already exist to detect faces in photos, but no model existed to determine face shapes. So we created one. We had to go online to get photos of people (mug shots were very convenient) to teach the model, but what we found is that since face shape identification can involve subtle differences in angles and curvature, determining the face shape by hand in order to teach the model was a difficult task for us. As a result, as much as we hate to admit it, our model is not as strong as it should be. I, however, am not too disappointed since we challenged ourselves and learned so much about AI and how to implement it in our projects.

Accomplishments that we're proud of

I am very proud that I was able to get much more familiar with iOS app development. This is a skill I can use in so many different circumstances and projects. Additionally, I'm very proud that we were able to integrate AI into our app. AI can sometimes seem like this daunting and mysterious thing, but being here at TAMUHack and having all of these resources and mentors made that hill much easier to climb.

What we learned

We learned so much about coding and motivation. I did more iOS development in 24 hours than I have over the past several months. Not only did that provide us with experience that we can take with us, but also a lesson that if we get in the right environment, we can make amazing projects from the ground up. We also learned just how accessible AI really is. It's such a blessing that big tech companies provide their elegant models and AI tools for us consumers to use. This allows us to implement AI in almost any project we want - something I would never have imagined if I had not worked on this project.

What's next for CalcuHandsome

We want to improve the model! Our app is nothing without a reliable model. We also want to add more features, such as using face shape to determine glasses and facial hair. Using this image we've created out of face shape, glasses, hairstyle, and facial hair, we can suggest outfits that would match this look. We can also scale to a point where companies can advertise their fashion products to our app users if those products match with our user's profile and analysis. This can help make the company profitable. Perhaps we can also have discounts from these advertisers exclusively on our app that you can't find in the department stores. And of course, let's make models for women! We see so many scalability opportunities here and hope this can be a go-to app when making fashion decisions.

Share this project:

Updates