Getting Information From Our Photos With Image Processing

Have you thought about how much information can your images or the ones online, have?

From knowing if your image is colorful enough, to know whether or not there are people in the picture, and if so, how many (and who they are).

Multiple faces were detected from people in the street

Artificial vision is, without a doubt, booming. Allowing progress that a few years ago for some has been unimaginable. And inside Artificial vision, there is this subfield of “image processing”.

Image processing is a powerful and hot topic nowadays mainly because it holds incredible power in automation:

• Doctors obtaining important data about their patients with Biomedical Imagining.

X-ray image of a chest with what seems to be pneumonia, next to the resulted enhanced image after applying image processing

• Designers or investigators reconstructing images, to gain better-detailed images that might have defected parts. Or even remove text or logos from these images.

Removing logos and text from images applying Image Processing algorithms with Python

• Developers or data scientists rescaling the resolution of images massively, before uploading them to a website or passing them to a Deep Learning model for training.

Rescaled resolution of an image, from 3000px to 800px

• Robots for recycling and cleaning our environment by identifying recycling waste items (glass bottles, plastic cartons, aluminum cans, etc)
• Factories that use programs to detect the shapes of objects in an image so they can know, for example, how many candies (or products) are in a table. Or count the number of points you have while playing domino!

Extracting information from domino’s tokens, like the total number of dots.

• Smart refrigerators that determine their contents, while providing this updated information to connected users.
• Apps that detect faces so you can, later on, apply filters, offer tagging friends, and almost anything else you can imagine!

Original image of me next to a resulted image with my face detected

As our human visual understanding of the world is reflected in our ability to make decisions through what we see, providing such visual understanding to computers would allow them the same power.

In this sense, images can contain as much information as we could imagine; information that can help us to advance in technologies for collective goods like diagnosing diseases faster, study space, or even create entertaining apps to spread joy around the world.

With Python, you can do many of these image manipulations seen above, with just a couple of lines!

I teach people how to do this in my course “Image Processing with Python” in Datacamp.

One of the exercises in the course: Learn to anonymize data in your photos automatically

Anyone can start practicing these techniques and it’s easier than you might think. Starting from the basics to more advanced, yet practical manipulation algorithms.

Even if you are a beginner, you can still get going, because on the platform there are other courses to get you where you want to be in the Data Science world! 🙌

You will be able to apply your knowledge to all differents domains that can use automation. Such as medical image analysis, retailing, space studies, and many more.

You can even try it with your friends' photos or with any of the images from growing online datasets like the ones found in Kaggle.

“The beautiful thing about learning is nobody can take it away from you.”

– B.B. King