Try This Today (TTT)

TensorFlow inception v3 / Android Application

Undergraduate Artificial Intelligence Course Group Project

Overview
‘Try This Today’ (TTT) is an artificial intelligence based Android app that made outwear suggestions to the users based on their closet inventory.

Purpose
People are having a hard time choosing their clothes based on the weather forecast. The goal of Try This Today (TTT) is to help users save time deciding on their clothes when preparing to go out. This application is providing to the users what kind of clothes to wear depending on the outside weather using artificial intelligence.

Application workflow

Three pages of application (`Main` page / `Add clothes` page / `My closet` page)
Two pages of application (`Long Sleeves` page / `Recommendation` page)
  1. Main: When you open the application, you can see the outside weather according to the location.
  2. Add clothes: Users can add their own clothes by clicking ‘Add Cloth’ button on the main page. After users took a picture of their own cloth, the trained model will classify the clothes. When the user press ‘Save’ button, then the cloth will be saved in the closet.
  3. My closet: Users can see what clothes they saved in this application by clicking the list of categories.
  4. Long sleeves: You can see the cloth that the user saved it before.
  5. Recommendation: When users click ‘Clothes Recommendation’ button on the main page, the application provides random three clothes for each category based on the weather.

How we built it
We trained the machine to classify the kind of clothes using TensorFlow inception version 3. Then, we referred to Tensorflow deep learning example to connect Android Studio and the model. To get the weather information, we used OpenWeather API. Also, we used Realm DB to save the clothes’ images.