Advanced level · Module 3 of 6 · 22 guided hours
Computer Vision Beyond Classification
Locate and outline objects with detection and segmentation models, evaluate them correctly and apply them to real tasks.
Learning objectives
- Train and evaluate object detection models
- Understand and apply image segmentation
- Design vision solutions for agriculture, health and infrastructure
Topics
- bounding boxes and annotation
- object detection models
- IoU and mean average precision
- semantic and instance segmentation
- counting and measuring
- application design
Lessons
-
Lesson 1 Object Detection
Find and label several objects in an image with bounding boxes.
bounding boxannotation toolsone-stage and two-stage detectorsnon-maximum suppression
-
Lesson 2 Evaluating Detection
Measure detection quality with intersection over union and mean average precision.
intersection over union (IoU)true and false detectionsaverage precisionmAP
-
Lesson 3 Segmentation
Classify every pixel to outline regions and objects precisely.
semantic segmentationinstance segmentationmaskspixel-level metrics
-
Lesson 4 Designing Vision Applications
Turn vision capabilities into useful applications for real users and conditions.
task choicefield conditionsuser workflowhuman confirmation
Practical activity
Fine-tune a pre-trained detector to find and count pest insects on sticky-trap photos from a provided open or synthetic data set, and evaluate it with IoU and mAP.
Expected outcomes
- Working detection and segmentation skills
- Correct evaluation of localisation tasks
Assessment
Detection practical marked on annotation quality, mAP results and error analysis.
Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.