Virtual Closet.
An AI-powered app to help manage your closet and pick out clothes faster.
Roles: Designer (Group project) with Mollie Dowst and Sara Berndt
Timeline: 2 weeks
Context: Coursework
Motivation: Getting dressed in the morning takes a lot of time and mental energy at the exact worst time to require it - right after you wake up. It is easy to wear the wrong outfit for the occasion, it is hard to make a decision, and you might end up too hot or too cold. There have been many attempts at virtual closets or outfit planning applications, but they do not make recommendations and have been described as tacky by some former users. Our virtual closet will tackle all of these issues and more, while delivering a great user experience.
We began by brainstorming the difficult parts of out daily lives such as remembering to take pills or remembering garbage day. Then, we began sketching our ideas and finally prototyping and and planning our automation strategies and task analysis.
Tools: Adobe XD, Adobe Photoshop, Adobe Illustrator, Excel
System Constraints
Although this application employs powerful AI, there are still some constraints that determine if the user or the system completes a task. As of now, our system is purely virtual and requires the user to complete the physical interactions with the clothes.
In the future, our system might pair with a physical product (such as a smart closet) which would be especially useful for a user in a wheelchair.
The system can’t
Dress the user
Wash and fold the clothes
Organize clothes in the closet
Make a final decision about the outfit that is worn
The system can
Select clothes that user would like to wear
Know what is available in the user’s closet
Determine outfits based on weather, events, schedule
Learn from previous user choices
Task Analysis
Task Analysis is the process of breaking a process down into a series of smaller tasks that can be understood to find better ways of completing said process. We completed a task analysis to understand how both the user alone and the user with our product might complete the process of getting dressed in the morning.
One notable part of our task analysis is the automation. By the system determining what the user should wear over time, the system becomes better at selecting outfits for the user which further offloads thinking, information gathering, and time from the user to the system.
To see the full Task Analysis, click here
Task Allocation is a way to determine whether the user or the system completes a task. In general, the process of adding clothes to the closet is completed by the user, but the process of selecting the day-to-day outfits based on the user’s criteria (events, weather, etc.) is completed by the system.
Automation Strategy
Automation Strategies describe how much automation the system has at a certain step. This range is from no automation to entirely automated.
Offers no assistance; the human must do it all
Suggests a complete set of alternatives to do the task
Narrows the set of alternatives to do the task
Recommends one alternative to do the task
Executes the recommendation if the human approves
Allows the human a restricted time to veto before automatic execution
Executes automatically, then necessarily informs the human
Executes automatically, then informs the human only if asked
Informs the human only if it decides to
Selects the method, executes the task, and ignores the human
Overall, our system either is mostly automated, or is hardly automated to support the user but the user completes the task.
Flow Chart
Flow Charts demonstrate key information necessary to complete tasks, showing how the system will operate. Clearly, the original task analysis with no automation support required much more information and steps from the user.
User Interface
Welcome.
1. Welcome screen
The first screen a new user sees.
2. Account set up
The user can sign in or create an account.
AI Calibration.
3. Body Type
The user can set his or her body type to help the AI shop for you or pick out clothes.
4. Sunny.
These screens allow the AI to calibrate outfits based on what the user would wear in certain weather situations.
5. Snowing.
There are four initial AI calibration screens that guide the user through as the onboarding process.
6. Windy.
The user is presented with some temperatures and clothes options to select between two demonstration outfits.
Add Clothes to your Virtual Closet.
7. Add clothes
Here the user can either add their clothes to the Virtual Closet or skip it for now.
8. Adding Instructions
The app explains to the user how to add the clothes.
9. Successful Scan
The user scans an item and if it is successfully added then the app shows a green check.
10. Failed Scan
The user did not position the shirt correctly so the app was unable to add the shirt to the virtual closet.
11. Successful Scan
The user scans an item and if it is successfully added then the app shows a green check.
12. Failed Scan
The user did not position the shirt correctly so the app was unable to add the shirt to the virtual closet.
Tour, Home Screen, and Notifications
13. Tour
The app offers the user a tour of the app’s functionality after completing the onboarding process.
14. Home Screen
The home screen displays the current weather that may be useful in determining what the user should wear as well as a selected outfit based on the current weather. The user can swipe to select other clothes. Also, the user can ‘remix’ the outfit.
15. Morning Notification
The app notifies the user to see the selected outfit each day based on the weather and the user’s typical choices.
