Ways of Seeing
Seeing is for Humans, and ComputersTsuhan Chen, Electrical and Computer Engineering
Images are represented by pixels, which are just numbers with values. We analyze these numbers and detect what’s going on in a picture.
Computing the Visual
We have all experienced the evolution of computer vision: photography enhancement, making the photos we take look better (making them sharper, improving the color, or removing granular noise); industry inspection, using cameras on industrial production lines to inspect parts as they are made; video surveillance, using cameras to analyze people going through airport security; even cameras that help drivers see better, like detecting a child in the driveway or an upcoming stop sign. And the movie Avatar has dazzled us. Our research has progressed steadily through these areas of computer vision technology.
Now, to this repertoire, we add research in the social sciences. We study pictures people take and share among themselves on social networks—analyzing the data they reveal. Our research is to understand the social relationships among people and how they develop.
I use the term “visual computing” to represent everything we do that utilizes computing to process visual data: graphics, pattern recognition, computer vision, and more.
What Your Facebook Photos Tell Us
If you are an outdoor person, you are likely to upload pictures with trees, mountains, rivers, or national park scenes in the background to Facebook. If you are an introverted person who spends time mostly in an office, you are more likely to post photos with computers or pictures on the wall in the background. By analyzing these kinds of images posted on Facebook, we know a lot about people.
To read the article in its entirety view below or download the full article.