"Father of YOLO" Joseph Redmon announced his withdrawal from the field of computer vision! This news that just appeared really surprised the artificial intelligence community.
On the social network, the inventor of the well-known YOLO AI algorithm suddenly stated yesterday: Out of moral considerations, he decided to stop all research on computer vision.
In the field of AI, this is the first time.
Speaking of YOLO, I believe that every computer vision practitioner is familiar with it. It is a very commonly used target detection algorithm. The task is to find out the target of our interest in the image, determine its size and position, and identify which object it is. From autonomous driving to face recognition, many common tasks in daily life are inseparable from this algorithm.
The YOLO model was first proposed by Joseph Redmon et al. in 2015 and revised in several subsequent papers.
Faster R-CNN and its improved Mask R-CNN have achieved good results in tasks such as instance segmentation, target detection, and human key point detection, but are usually slow. The innovation of YOLO is that it proposes one-stage, that is, target positioning and target recognition are completed in one step, which is a veritable "You Only Look Once".
Since YOLO only uses a single network, it can be directly optimized end-to-end on detection performance, enabling the base YOLO model to process images in real time at 45 frames per second. A small-scale version of YOLO - Fast YOLO can achieve a processing speed of 155 frames per second.
YOLO has an amazing speed, but also has a flaw that stops people: it is not good at small target detection. In order to make up for this shortcoming, in 2018, Redmon et al released YOLO v3. This new version maintains the speed advantage of YOLO, improves the accuracy of the model, especially strengthens the recognition of small targets and overlapping occluded targets, and complements the shortcomings of YOLO. It is the current target detection network with balanced speed and accuracy.
The researchers' outlook for the next version of YOLO mainly lies in three aspects: higher recognition accuracy, more extensive real-time monitoring, and a lighter model. On GitHub, the answer to the question of when the v4 version will be released has always been "coming soon".
For a long time, Joseph Redmon has been engaged in computer vision research with Allen School professor Ali Farhadi. He is the winner of the 2018 Google Ph.D. Scholarship for his contribution to "creating faster, better, and more useful computer vision application tools".
Share some of the artificial intelligence learning materials I have compiled for you for free. It has been compiled for a long time and is very comprehensive. Including some common AI framework actual combat videos, image recognition, YOLO, OpenCV, NLQ, machine learning, pytorch, computer vision, deep learning and neural network and other videos, courseware source code, well-known domestic and foreign elite resources, popular AI papers, industry reports, etc.
In order to better systematically learn AI, it is recommended that you collect a copy.
The following are some screenshots, and the free download method is attached at the end of the article.
1. Artificial intelligence must-read books
2. Artificial intelligence free video courses and projects
3. Collection of Papers on Artificial Intelligence
4. Artificial intelligence industry report
To learn artificial intelligence well, you need to read more books, do more hands-on work, and practice more. If you want to improve your level, you must learn to calm down and learn systematically slowly, so that you can gain something in the end.