Demotion

The Basic GUI
Happiness Blur Test
Neutral Blur Test
Suprise Blur Test

Inspiration

We were interested in using the Microsoft Emotion API within a program to identify and modify an image. It was created to prevent a single person from ruining the mood of a group photo.

What it does

Blurs all faces within a photo that are showing the selected emotion (neutral, happiness, sadness, contempt, disgust, anger, surprise, or fear). The photo can be imported from the web or uploaded from a file directory. The blurred photo can then be saved to the user's computer.

How we built it

We used the Emotion API from Microsoft Cognitive Services to determine the emotions of each person within an image and the locations of their faces. A Gaussian Blur was then applied to each face that matched the selected emotion, and a new image was created.

Challenges we ran into

Importing the Microsoft Azure API, learning how to use Python, Problems with the pip package and the PIL module, and converting between Python 2 and 3.

Accomplishments that we're proud of

Successfully implementing the Microsoft API and extracting data from it, Learning the syntax for Python in a day, creating a working image modification program.

What we learned

How to implement an API into a program, how to use Python, how to implement a Gaussian Blur onto an image, how to use JSON, how to create a GUI in Python, basic Graphic Design.

What's next for Demotion

GUI improvements, ability to erase individual people from a photo, option to type the emotion you would like to remove.

Built With

api
azure
cognitive
microsoft
python

Submitted to

MHacks X

Created by

I worked on implementing the GUI for the project. This included adding the radio buttons, text boxes, and labels. Additionally, I created the variables and logic that allowed for communication between the GUI and the back-end. This included passing the URL/File name, and the selected emotion to be removed.

Matthew Hermann
I worked on implementing the Microsoft Emotion APIs and filtering the data to find the emotions and corresponding face locations that we wanted. I also worked on connecting the various parts of the project together. Additionally, I handled reading files and saving files on the user's computer.

Nicholas Kroetsch
I worked on installing and running the pip package and PIL module required to use the Gaussian Blur. Further, I tested the code while it was being developed, and I also connected the parts of the project with Nicholas. I learned a lot about using the Mac Terminal with Python Files!

Ishpreet Kohli

Updates

Matthew Hermann started this project — Sep 23, 2017 11:36 PM EDT

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.