Scribd
Upload
14 views4 pages

YOLO (You Only Look Once)

YOLO (You Only Look Once) is a real-time object detection algorithm that utilizes neural networks for high-speed and accurate object identification and localization. It operates as a regression problem and employs techniques such as residual blocks, bounding box regression, and intersection over union for effective detection. The evolution from YOLOv1 to YOLOv2 introduced improvements like batch normalization, higher resolution classifiers, and anchor boxes, enhancing its ability to detect smaller objects with greater precision.

Uploaded by

sathish
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
14 views4 pages

YOLO (You Only Look Once)

YOLO (You Only Look Once) is a real-time object detection algorithm that utilizes neural networks for high-speed and accurate object identification and localization. It operates as a regression problem and employs techniques such as residual blocks, bounding box regression, and intersection over union for effective detection. The evolution from YOLOv1 to YOLOv2 introduced improvements like batch normalization, higher resolution classifiers, and anchor boxes, enhancing its ability to detect smaller objects with greater precision.

Uploaded by

sathish
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 4

YOLO (You Only Look Once)

An algorithm that uses neural networks for real-time object detection.


It takes less time to detect the object with high accuracy.

Two basic things: -


• To identify objects in an image like human beings, animals, cars etc..
• To identify the exact location of an object in the image.

Object detection in YOLO is done as a regression problem and provides the class probabilities of the detected
image.
YOLO is important because it has three main things (Speed, High accuracy, and Learning capabilities)

1. The algorithm improves speed of detection, can predict objects in real-time.


2. YOLO is a predictive techniques that provides accurate results with minimal background errors.
3. The algorithm has excellent learning capabilities that enable to learn the representations of objects.

How YOLO algorithm works?


Mainly three techniques (Residual blocks, Bounding box regression, Intersection over union)

Residual blocks: - Image will be divided into various grids. Each grid has a dimension of S*S.

Bounding box regression: -


• A bounding box is an outline that highlights an object in an image.
• Bounding box consists of width and height.
• Example of class (cars, bikes, peoples, animals etc..) represents c.
• Bounding center (bx, by)
• YOLO uses a single bounding box regression to predict the height width, center and class of the objects.

Intersection over union: - Intersection over Union is an evaluation metric used to measure the accuracy of an
object detector on a particular dataset.
Difference between YOLOv1, YOLOv2 or 9000
YOLOv1: -
• V1 uses darknet frame for training image dataset.
• Darknet is an architecture, used as a neural network framework for training YOLO.
• YOLO v1 could not find small objects in an image as if they appeared in the cluster, is restricted.
• This architecture found difficulty in generalization of objects if the image is of other dimensions different from the
trained image.
• The major issue is localization of objects in the input image.

YOLO v2 or 9000: -
Batch normalization: -
• Batch normalization decreases the shift in unit value in the hidden layer and by doing so it improves the stability
of the neural network.
• Its precision has been increased by 2%. Also helps model to regularize and overfitting has been reduced overall.
Higher resolution classifier: -
The input size in YOLO v2 has been increased from 224*224 to 448*448, increase in the precision up to 4%.
Anchor boxes: -
• Most notable changes is visible in YOLOv2, introduction anchor boxes.
• V2 does classification and prediction on single frame work.
• Anchor box is responsible for predicting bounding box.
Darknet architecture: -
• YOLOv2 uses darknet 19 architecture, that has 19 convolution layers, 5 max pooling layers and a softmax
layer for classification of objects.
• YOLO v2 is better, faster, and stronger.
• YOLO v2 has seen a great improvement in detecting smaller objects with much more accuracy which it
lacked in its previous version.

You might also like