Image Processing: Convolution, Kernel, Aliasing.
Sometimes, I wonder how my monitor works and why image data looks like that because I know the signal is continuous. If you are interested in computer vision or image classification based on deep learning, this post is necessary for you. Knowing how images can be shown means you can understand your data more precisely. Quantitative data, gene expression, stock data, and e.t.c, needs proper domain knowledge to interpret the semantics inside. It is the same with image data. You should know how it is represented, what is needed to be translated from the signal to your monitor, and what function is needed. I will explain this with intuition. if you want to exact formula or process, then you recommend you to watch MIT OpenCourseWare or other MOOCs because electronic engineering guys study this for 1~2 semesters.
The picture is taken at Suwon, South Korea where my bachelor university is located in. Do you like this picture? How did the camera capture the light? As I mentioned in previous posts, light can have wave and particle properties at the same time. In this post, we will consider light is a wave. What the camera did is sample the intensity of the wave at some point and quantize it. You can show the quantized value like this
