Team 1 - Fashion Fusion

• 

  UI_page1

• 

  UI_page2

• 

  UI_page3

• 

  UI_page4

• 

  Examples from our Training Data Set

• 

  ML Model

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

Inspiration

Based on our research, we learned that most Fashion Designers actually get their inspiration by consulting physical art. We used this concept to create a program that helps users, who aren’t fashion designers, do the same for their own outfits.

What it does

We created at app that produces customized outfit inspirations based on an artistic image that the user inputs. Essentially, it fuses the user’s artistic choices and interests with fashion and outfits, hence the name Fashion Fusion.

How we built it

We collected our own data, composed of 553 images of a variety of themes to train our ML algorithm. We also put together around 80 outfit designs for our project. We trained the model to identify and categorize images based on their color spectrum, the tone, elements and settings. After the model was trained we tested the model with 50+ images, getting accurate predictions for 98.01% of the rounds. The code also implements a user interaction section where the user can insert their own image and the trained model will analyze the image, make a prediction and provide the correlated results. We created this on google colab, using Python and Tensorflow.

Challenges we ran into

Because we have little experience with machine learning, it was extremely difficult to make the machine work and understand how it worked as well. Thus, we relied on mentors and YouTube tutorials to aid us with the coding process.

Accomplishments that we're proud of

We're proud that the machine learning algorithm can accurately classify an image based on its aesthetic.

What we learned

We learned how to work together on a coding project. We also learned how to use machine learning algorithms on Google Colab.

What's next for Fashion Fusion

We plan to add more categories of aesthetics, such as different cultural styles, modern aesthetics, and traditional aesthetics. We plan to train ML code program to recognize these aesthetics to integrate it with the app. We'll also add more categories of clothing, such as business attire, sportswear, and swimwear. Finally, we'll finish working on the front-end and connect front-end to back-end.

Built With

• canva
• google-colab
• python
• tenserflow