Using YOLOv3 in Keras for identifying objects is one of the foundational tasks of machine learning.
The “You Only Look Once” algorithm is a popular one for object detection, since in real life, you really only get one shot to figure out what something is.
Humans don’t get the luxury of multiple perspectives and time-delayed training sets, so YOLO is more real-life than other neural net algorithms.
For the tutorial I found, you’ll use the darknet source, as it is the one referenced by the author.
The tutorial will use Experiencor as well.
The steps in the tutorial are as follows:
1. YOLO for Object DetectionML.AI
2. Experiencor YOLO3 Project
3. Object Detection With YOLOv3
Here are the references from the tutorial, in case you wish to dive deeper in the foundations of this machine learning methodology:
You Only Look Once: Unified, Real-Time Object Detection, 2015.
YOLO9000: Better, Faster, Stronger, 2016.
YOLOv3: An Incremental Improvement, 2018.
YOLO: Real-Time Object Detection, Homepage.
Official DarkNet and YOLO Source Code, GitHub.
Official YOLO: Real Time Object Detection.
Huynh Ngoc Anh, experiencor, Home Page.
Other YOLO for Keras Projects
There are a number of steps involved in this YOLOv3 approach, including model creation, prediction, and measuring and adjusting results.
To get started, proceed to machine learnings to begin.
For more information about machine learning, click here.