Hello trend,
Photography literally means "writing with light." Needless to say, the quality of this "writing" suffers when there isn't enough light. Today, we are sharing a post on improving the quality of images taken under low light conditions.
If you are taking a picture yourself, you can fix this problem by using a flash, or by increasing the size of the aperture (f-stop), or by keeping the camera aperture open for a longer time assuming you have a largely static scene.
But if you do not control the picture, the problem becomes tough to fix. Fortunately, not all is lost. You will learn it is possible to use certain statistics of the image called the dark channel prior and bright channel prior to recover information from a single badly lit image.
https://learnopencv.com/improving-illumination-in-night-time-images/
We are sharing code in both C++ and Python at the link below.
https://github.com/spmallick/learnopencv/tree/master/Improving-Illumination-in-Night-Time-Images
First Principles of Computer Vision
When I was in grad school, I read every paper of some of my favorite researchers. Dr. Shree Nayar of Columbia University was on the top of that list. His papers on physics-based vision illuminated new concepts with clarity and elegance.
I was pleasantly surprised when he recently reached out to me via linkedin and informed me about his free lecture series titled "First Principles of Computer Vision." I went through several videos in the series. I wish I had access to these video lectures when I was in grad school.
You recommend you check them out. This is the kind of free material you should go and consume before you pay for courses.
When you are ready for a hands-on course experience with lots of practical examples, I recommend you take the Official AI courses by OpenCV.org. Click the button below to join the waitlist - the wait time is about two weeks. People on the waitlist get priority, and also a chance to buy the courses at a 20% discount when the enrollment opens.
| JOIN THE WAITLIST |
Satya
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.