smoovacado

Inspiration

We loved watching CSI as kids, and we wanted to make the infamous "enhance" a reality.

What it does

Smoovacado uses Gaussian blurs and several edge detection methods, namely Sobel and Canny to blur and sharpen the image. We believe that Smoovacado could be used to help security cameras identify faces.

How we built it

We started learning about how blurs on images worked from this video. After learning about the different blurs and filters, we looked at a few different libraries: SciPy, NumPy, matplotlib, scikit-image, Python Imaging Library. We converted images to grayscale and ran filters on them in different orders until we found an order that led to good results. We then focused our attention towards building a simple and clean web interface.

Challenges we ran into

We had immense trouble with figuring out how to have the user upload files to our site and pass that file to our python image processor.

Accomplishments that we're proud of

None of our team members had significant experience in python or back-end development, so learning it in a weekend was a pretty fun and challenging experience.

What we learned

Back-end development is very important and should be a focus from the get-go. Python is very versatile

What's next for Smoovacado

Instead of setting arbitrary weights across the code, use machine learning algorithms to optimize and select the proper weights for each image. Completing the website and adding drag-and-drop functionality for adding images. Apply Smoovacado to videos.

Built With

• artificial-intelligence
• azure
• computer-vision
• css
• html5
• javascript
• python