This week you'll learn about Achieving Optimal Speed and Accuracy in Object Detection (YOLOv4).

Image

You are probably wondering what else is there to work on with YOLOv3 checking almost all boxes and bringing in monumental upgrades over its predecessors. But if you remember, YOLOv3, even with all of its upgrades, did fall short in one of its mean average precision (mAP) evaluations. 

So does YOLOv4 get better than its predecessor in this angle? Was YOLOv3 too good to be toppled? Or did YOLOv4 focus on other ways of bettering itself?

Today, we will learn about YOLOv4 and the various ways it can help you reach optimal efficiency and accuracy in object detection. 

The big picture: YOLOv4 utilizes several strategies to maintain not only its high accuracy but also its high speed and frame yields. These are known as the bag of freebies and the bag of specials. Both of these aid YOLOv4 significantly in becoming an all-around better object detector.

How it works: The YOLOv4 brings a lot of updates in itself, like the CSPDarknet53 backbone. The bag of freebies is the term given to techniques that help change the training strategy giving better accuracy. Techniques like data augmentation, semantic distribution bias in datasets, etc., are used in the training stages of YOLOv4. The bag of specials is termed as plugins and modules that increase the inference cost but give a massive boost to accuracy. These consist of the "Mish" activation function, Spatial Attention modules, etc. 

Our thoughts: Learning about YOLOv4 is extremely interesting, but understanding the thought process between the bag of freebies and specials gives us great insight into how the goal of reaching optimization is realized. All these concepts cover a wide area of interest and can help you in many other endeavors. 

Yes, but: The sheer number of concepts might feel intimidating to you!

Stay smart: Take your time, and learn at your own pace. But be sure you see it through. We hope this blog stands as a massive leap for you not only in object detection but also in other deep learning domains. 

Click here to read the full tutorial

Solve Your CV/DL problem this week (or weekend) with our Working Code

You can instantly access all of the code for Achieving Optimal Speed and Accuracy in Object Detection (YOLOv4) by joining PyImageSearch University. Get working code to

  1. Finish your project this weekend with our code
  2. Solve your thorniest coding problems at work this week to show off your expertise
  3. Publish groundbreaking research without multiple tries at coding the hard parts

Guaranteed Results: If you haven't accomplished your CV/DL goals, let us know within 30 days and get a full refund.

I want the code


The PyImageSearch Team
Image


P.S. If you're interested in learning how to successfully apply deep learning to your own projects, I would recommend reading my book, Deep Learning for Computer Vision with Python.

Inside the book you'll find:
  • Super-practical walkthroughs that present solutions to actual real-world image classification (ResNet, VGG, etc.), object detection (Faster R-CNN, SSDs, RetinaNet, etc.), and segmentation (Mask R-CNN) problems
  • Hands on tutorials (with lots of code) that show you not only the algorithms behind deep learning for computer vision but their implementations as well.
  • A no-nonsense teaching style that is guaranteed to help you master deep learning for image understanding and visual recognition
If you're interested in learning more about the book, I'd be happy to send you a PDF containing the Table of Contents and a few sample chapters:

Click here to grab the PDF of sample chapters and Table of Contents

After clicking the above link, you'll receive a separate email with the PDF in a few short moments.