Hello trend,
This is Satya Mallick from LearnOpenCV.com.
We have an informative new post on Non Maximum Suppression with PyTorch today.
Imagine you have trained a car detector. If trained properly it will create a bounding box around all cars in an image. Now, if you move the detected bounding box by one pixel in any direction, it is still a valid picture of a car. The detector may also select bounding boxes that do not cover the entire car.
So, it will end up producing more than one bounding box for the same object unless we do some post-processing to filter out these multiple detections for the same object. The class of algorithms for achieving this filtering is called Non Maximum Suppression.
In today's post, we go over the nuances of the problem and share an implementation
https://learnopencv.com/non-maximum-suppression-theory-and-implementation-in-pytorch/
The code in PyTorch for the above post is at
https://github.com/spmallick/learnopencv/tree/master/Non-Maximum-Suppression
OpenCV for Beginners will be FREE for Professors
Good teachers can transform societies! It is easy to forget that our teachers, too, have to upgrade their skills and get trained in new and fast-changing technologies like computer vision and AI.
To express our gratitude toward outstanding teachers, we are making our course, OpenCV for Beginners, free for professors worldwide. All they need to do is fill out this form
https://opencv.org/opencv-for-university/
The course will start in Sep 2021.
For everyone else, consider backing our Kickstarter campaign for courses that start with the basics to take you to mastery in computer vision and AI.
YOLO v3, v4, v5, Oh My! : OpenCV Weekly Webinar in 15 minutes
In today's OpenCV Weekly Webinar we will go over the various flavors of the popular object detector YOLO
Topic : YOLO v3, v4, v5, Oh My!
Time : 9 AM Pacific Time, June 3, 2021
Hosts : Satya Mallick (CEO, OpenCV.org), Joseph Nelson (CEO, Roboflow), and Phil Nelson (Content Manager, OpenCV.org)
Free registration : Zoom Registration Link
Also, check out last week's episode where I had a very entertaining discussion with Vikas Reddy, cofounder of Occipital and LightTwist, about 3D Vision For Virtual Video Production.
Cheers!
Satya
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.