Why computer vision?

As you see, I claim I am good at computer vision. Here I am going to account for why I choose computer vision as a way to go.

I am a universal person during my student period, that is, I almost interest in every courses in Mathematics and Statistics. After I graduate from Statistics department, I always think how can I apply these knowledge in my work? It is challenging since math is a really abstract science. How group theory works? How geometry works? How complex analysis works? Except some straightforward courses such like machine learning or cryptograph, the other math branches seem too far to have a real application.

However, this idea is very wrong. Indeed, there are bunch of math problem waiting for us. Such like a manufacturing process may link to the job-shop scheduling problem, and the job-shop scheduling problem is usually modelled by a DAG and Markov Decision Process. It’s so excited! The real world problem can be translated to my familiar math language.

Besides of job-shop scheduling problem. There are still plenty of math applications in the real world. I have no option but to say, computer vision is the richest field which includes most intricate math and statistical applications. Computer vision not only refers to digital image process, all of the sensor to help computer understand the world can be interpreted as one of computer vision. For example, a LiDar scan or an Optical Flow can be also interpreted as a computer vision. To extract the information from such scanning, we need both vertical and horizontal knowledge of it. We have projective geometry to help us build a relationship between real world and picture. We have Fourier Series to interpret the frequency of image. We have 3D rotation group to distinguish sensor from different angle.

The other item I want to highlight is, computer vision is highly related to the programming skill. For example, to identify if a point lies in a rectangle is quite easy. However, we can only read such the information from computer

Coordinate X Y
North West -2.7 9.4
North East 2.7 12.5
South East 6.4 6.1
South West 1 3

To determine if a point (3.4, 1.0) lies in the rectangle, it is not straightforward, right? It’s rely on the programing skill to determine it even a simplest problem.

There are too many materials to say. Even though some of the math are challenging for me! For now, you can see why I choose computer vision as my career due to its challenging! Hope this article would also inspire you in some sense.

2020

Contour Fitting : Parametric Curve

2 minute read

I want to talk about a general contour fitting approach. Before introduction, I would like to clarify the difference between contour fitting and curve fitti...

Why computer vision?

2 minute read

As you see, I claim I am good at computer vision. Here I am going to account for why I choose computer vision as a way to go.

First post!

less than 1 minute read

Hello world, this is my first blog post.

Back to top ↑