liveBook · Manning

Intersection over Union

Also called IoU

Intersection over Union is the ratio of the intersection area or volume of two regions to the area or volume of their union.

How 2 Manning books explain it

Math and Architectures of Deep Learning

A bounding-box overlap metric used to match proposed regions to ground-truth boxes.

How it is used
Matches anchors to ground-truth boxes for RPN labels and filters duplicate proposals during non-maximal suppression.
How deep it goes
Defines overlap over union, explains its geometric interpretation, and gives threshold-based anchor-labeling and PyTorch computation procedures.
Examples
Uses overlapping bounding-box geometry, GT-to-anchor score matrices, and torchvision.ops.box_iou and nms calls.
What you can do after
Apply it via torchvision.ops to assign anchor labels and suppress overlapping region proposals.

Read how Math and Architectures of Deep Learning explains Intersection over Union →

Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI

A measure of overlap between a predicted region and a ground-truth region, calculated as their intersection area divided by their union area.

How it is used
Measures bounding-box localization accuracy and agreement, and supports selection of detections for human review.
How deep it goes
Defines area overlap and works through annotation, agreement, sampling, and chance-adjustment uses without deriving the metric.
Examples
Illustrates overlapping bounding boxes, multiple annotator boxes, and a chance adjustment for an object covering 10% of an image.
What you can do after
Evaluate bounding-box localization and annotator agreement, and identify uncertain detections for review.

Read how Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI explains Intersection over Union →

Also covered in

What to know first

Related concepts

Questions about Intersection over Union

What is Intersection over Union?

Intersection over Union is the ratio of the intersection area or volume of two regions to the area or volume of their union.

Which Manning books explain Intersection over Union?

Math and Architectures of Deep Learning and Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI.

What should I know before learning Intersection over Union?

Manning books that teach Intersection over Union usually introduce bounding box, convolutional neural network, feature map and Region Proposal Network first.

What is the difference between Intersection over Union and f-score?

Intersection over Union is the ratio of the intersection area or volume of two regions to the area or volume of their union. An F-score is a classification metric that combines precision and recall through a weighted harmonic mean; F1 weights them equally.

What is Intersection over Union also called?

Intersection over Union is also called IoU.

These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.