Jump-Scared

Inspiration

I've always been intrigued by the idea of blending technology with traditional festivities. With the increasing advancements in facial recognition and machine learning, I wanted to create something fun, relevant, and a little spooky for the Halloween season.

What it does

Jump-Scared is an innovative emotion detection project. It continuously monitors the user's facial expressions and, when it detects a neutral face for over 15 seconds, it surprises the user with a jump-scare in the form of a spooky GIF and a blood-curling sound!

How I built it

Using the power of Python, I integrated OpenCV and Dlib for facial detection and DeepFace for emotion recognition. The GUI is powered by Tkinter to display the scary GIF, and the sound is managed through Python's subprocess module.

Challenges I ran into

• Handling multiple monitors and displaying the GIF fullscreen on all connected displays was especially tricky, for anyone who has used python for automation before, you may know that python isn't exactly friendly with external displays.
• Ensuring that the GIF animation plays smoothly without freezing - my computer isn't the most powerful, so I learnt a lot about optimizing a system for the best possible performance.
• Balancing system resources, as opening multiple instances of jump-scares could be resource-intensive (accidentally crashed my laptop a few times!).

Accomplishments that I'm proud of

• Figuring out emotion detection and facial detection.
• Creating multiple scripts that work together quite well, each on their own subprocess
• Successfully integrating multiple libraries for a seamless user experience.
• Creating an interactive and fun application that can genuinely surprise (and perhaps scare!) someone.

What I learnt

• The intricacies of facial emotion detection.
• Efficiently managing and optimizing system resources for a smoother experience.
• The importance of user feedback in refining and enhancing the project (When I told my family about a hackathon project, they were more than happy to help me test my code-but I may have forgotten to tell them what exactly the project was!).

What's next for Jump-Scared

• Introducing more diverse reactions based on different emotions detected.
• A Scare-O-Meter. Quantifying the level of fright a user feels with facial-based emotion detection!
• Enhancing the GUI for a more interactive and immersive user experience.
• Possibly integrating with IoT devices for a holistic Halloween setup – imagine lights flickering in sync with the jump-scare!

Built With

• emotion-recognition
• facial-recognition
• python
• subproccess