book

OpenCV with Python Blueprints

by Michael Beyeler, Michael Beyeler (USD)

October 2015

Intermediate to advanced

230 pages

5h 10m

English

Packt Publishing

Read now

Unlock full access

Support files, eBooks, discount offers, and moreWhy subscribe?Free access for Packt account holders
What this book covers

Downloading the example code
ErrataPiracyQuestions
Planning the app
Implementing dodging and burning in OpenCVPencil sketch transformation
Color manipulation via curve shiftingImplementing a curve filter by using lookup tablesDesigning the warming/cooling effect
Using a bilateral filter for edge-aware smoothingDetecting and emphasizing prominent edgesCombining colors and outlines to produce a cartoon
Running the appThe GUI base classThe GUI constructorHandling video streamsA basic GUI layoutA custom filter layout
Planning the app
Accessing the Kinect 3D sensorRunning the appThe Kinect GUI
Finding the most prominent depth of the image center regionApplying morphological closing to smoothen the segmentation maskFinding connected components in a segmentation mask
Determining the contour of the segmented hand regionFinding the convex hull of a contour areaFinding the convexity defects of a convex hull
Distinguishing between different causes of convexity defectsClassifying hand gestures based on the number of extended fingers
Tasks performed by the app
Running the appThe FeatureMatching GUI
Feature detectionDetecting features in an image with SURF
Matching features across images with FLANNThe ratio test for outlier removalVisualizing feature matchesHomography estimationWarping the image
Early outlier detection and rejection
Planning the app
The pinhole camera modelEstimating the intrinsic camera parametersThe camera calibration GUIInitializing the algorithmCollecting image and object pointsFinding the camera matrix
The main function routineThe SceneReconstruction3D class
Point matching using rich feature descriptorsPoint matching using optic flowFinding the camera matricesImage rectification
Planning the app
The main function routineThe Saliency classThe MultiObjectTracker class
Fourier analysisNatural scene statisticsGenerating a Saliency map with the spectral residual approachDetecting proto-objects in a scene
Automatically tracking all players on a soccer fieldExtracting bounding boxes for proto-objectsSetting up the necessary bookkeeping for mean-shift trackingTracking objects with the mean-shift algorithm
Planning the app
The training procedureThe testing procedureA classifier base class
Parsing the dataset
Common preprocessingGrayscale featuresColor spacesSpeeded Up Robust FeaturesHistogram of Oriented Gradients
Using SVMs for Multi-class classificationTraining the SVMTesting the SVMConfusion matrixAccuracyPrecisionRecall
Planning the app
Haar-based cascade classifiersPre-trained cascade classifiersUsing a pre-trained cascade classifierThe FaceDetector classDetecting faces in grayscale imagesPreprocessing detected faces
Assembling a training setRunning the screen captureThe GUI constructorThe GUI layoutProcessing the current frameAdding a training sample to the training setDumping the complete training set to a fileFeature extractionPreprocessing the datasetPrincipal component analysisMulti-layer perceptronsThe perceptronDeep architecturesAn MLP for facial expression recognitionTraining the MLPTesting the MLPRunning the script
Summary

Content preview from OpenCV with Python Blueprints

Face detection

OpenCV comes preinstalled with a range of sophisticated classifiers for general-purpose object detection. Perhaps, the most commonly known detector is the cascade of Haar-based feature detectors for face detection, which was invented by Paul Viola and Michael Jones.

Haar-based cascade classifiers

Every book on OpenCV should at least mention the Viola–Jones face detector. Invented in 2001, this cascade classifier disrupted the field of computer vision, as it finally allowed real-time face detection and face recognition.

The classifier is based on Haar-like features (similar to Haar basis functions), which sum up the pixel intensities in small regions of an image, as well as capture the difference between adjacent image regions. Some ...