Index

A

• abstract base class / The GUI base class
• abstract base class (ABC)
  • about / A classifier base class
• accuracy
  • about / Accuracy
• adaptive thresholding / Cartoonizing an image
• Alfheim dataset
  • URL / Automatically tracking all players on a soccer field
• app
  • summarizing / Putting it all together, Running the app
  • running / Running the app, Running the app, Running the app
  • GUI base class / The GUI base class
  • custom filter layout / A custom filter layout
  • setting up / Setting up the app, Setting up the app
  • Kinect depth sensor, accessing / Accessing the Kinect 3D sensor
  • Kinect GUI / The Kinect GUI
  • tasks performed / Tasks performed by the app
  • planning / Planning the app, Planning the app, Planning the app, Planning the app, Planning the app
  • FeatureMatching GUI / The FeatureMatching GUI
  • scripts / Planning the app
  • modules / Planning the app
  • prerequisites / Planning the app
  • implementing / Putting it all together, Putting it all together, Putting it all together
• app setup
  • about / Setting up the app, Setting up the app
  • main function routine / The main function routine, The main function routine
  • SceneReconstruction3D class / The SceneReconstruction3D class
  • Saliency class / The Saliency class
  • MultiObjectTracker class / The MultiObjectTracker class

B

• backpropagation
  • about / Multi-layer perceptrons, Deep architectures
• base class
  • about / A classifier base class
• BaseLayout class / A custom filter layout
• batch learning
  • about / Deep architectures
• bilateral filter
  • about / Cartoonizing an image
  • used, for edge-aware smoothing / Using a bilateral filter for edge-aware smoothing
• Binary Robust Independent Elementary Features (BRIEF) / Feature detection
• black-and-white pencil sketch
  • creating / Creating a black-and-white pencil sketch
  • dodging, in OpenCV / Implementing dodging and burning in OpenCV
  • pencil sketch transformation / Pencil sketch transformation
• bookkeeping
  • setting up, for mean-shift tracking / Setting up the necessary bookkeeping for mean-shift tracking
• bounding boxes
  • extracting, for proto-objects / Extracting bounding boxes for proto-objects
• burning / Creating a black-and-white pencil sketch

C

• C++ scripts, camera calibration
  • reference link / Estimating the intrinsic camera parameters
• camera calibration
  • about / Camera calibration, The pinhole camera model
  • pinhole camera model / The pinhole camera model
  • intrinsic camera parameters, estimating / Estimating the intrinsic camera parameters
• camera matrices
  • finding / Finding the camera matrices
• camera motion
  • estimating, from pair of images / Estimating the camera motion from a pair of images
• camera resectioning
  • about / Camera calibration
• Canny edge detection / Detecting and emphasizing prominent edges
• cascade of Haar-based feature detectors
  • about / Face detection
• classification
  • about / Supervised learning
• classifier base class
  • about / A classifier base class
• color spaces / Color spaces
• confusion matrix
  • about / Confusion matrix
• curve filter
  • about / Generating a warming/cooling filter
  • implementing, with lookup tables / Implementing a curve filter by using lookup tables
• cv2.findHomography function / Homography estimation
• cv2.pencilSketch function / Creating a black-and-white pencil sketch

D

• 3D point cloud visualization
  • about / 3D point cloud visualization
• de-meaning
  • about / Common preprocessing
• decision boundary
  • about / The training procedure
• designed image filter effects
  • custom filter layout / A custom filter layout
• detected faces
  • preprocessing / Preprocessing detected faces
• detectMultiScale function, options
  • minFeatureSize / Using a pre-trained cascade classifier
  • searchScaleFactor / Using a pre-trained cascade classifier
  • minNeighbors / Using a pre-trained cascade classifier
  • flags / Using a pre-trained cascade classifier
• Discrete Fourier Transform (DFT)
  • using / Fourier analysis
• distortion coefficients
  • about / The pinhole camera model
• dodging / Creating a black-and-white pencil sketch

E

• edge detection / Cartoonizing an image
• epipolar constraint
  • about / Reconstructing the scene
• epipolar geometry
  • about / Reconstructing the scene
• epipolar point
  • about / Reconstructing the scene
• epipole
  • about / Reconstructing the scene
• essential matrix
  • about / Estimating the camera motion from a pair of images

F

• face detection
  • about / Face detection
  • Haar-based cascade classifier / Haar-based cascade classifiers
  • pre-trained cascade classifiers / Pre-trained cascade classifiers
  • FaceDetector class / The FaceDetector class
• FaceDetector class
  • about / The FaceDetector class
• faces
  • detecting, in grayscale images / Detecting faces in grayscale images
• facial expression recognition
  • about / Facial expression recognition
  • training set, assembling / Assembling a training set
  • feature extraction / Feature extraction
  • Multi-Layer Perceptrons (MLPs) / Multi-layer perceptrons
• Facial Expression Recognition challenge
  • reference link / Summary
• false positives
  • about / Precision
• Fast Library for Approximate Nearest Neighbors (FLANN)
  • about / Tasks performed by the app, Feature matching
  • used, for feature matching across images / Matching features across images with FLANN
• feature detection algorithms, OpenCV
  • Harris corner detection / Feature detection
  • Shi-Tomasi corner detection / Feature detection
  • Scale-Invariant Feature Transform (SIFT) / Feature detection
  • Speeded-Up Robust Features (SURF) / Feature detection
• feature engineering
  • about / The training procedure
• feature extraction
  • about / Tasks performed by the app, Feature extraction, Feature extraction
  • feature detection / Feature detection
  • common preprocessing / Common preprocessing
  • grayscale features / Grayscale features
  • color spaces / Color spaces
  • Speeded-Up Robust Features (SURF) / Speeded Up Robust Features
  • Histogram of Oriented Gradients (HOG) / Histogram of Oriented Gradients
• feature extraction, facial expression recognition
  • about / Feature extraction
  • dataset, preprocessing / Preprocessing the dataset
• feature matches
  • visualizing / Visualizing feature matches
• feature matching
  • about / Tasks performed by the app, Feature matching
  • across images, with FLANN / Matching features across images with FLANN
  • outlier removal, ratio test / The ratio test for outlier removal
  • homography estimation / Homography estimation
  • image, warping / Warping the image
• FeatureMatching GUI / The FeatureMatching GUI
• feature selection
  • about / The training procedure
• Features from Accelerated Segment Test (FAST) / Feature detection
• feature tracking
  • outlier detection / Tasks performed by the app, Early outlier detection and rejection
  • outlier rejection / Tasks performed by the app, Early outlier detection and rejection
  • about / Feature tracking
• feed-forward neural network
  • about / Deep architectures
• first principal component
  • about / Principal component analysis
• focal length
  • about / The pinhole camera model
• Fourier analysis
  • about / Fourier analysis
• Fourier transform
  • about / Fourier analysis
• frequency domain
  • about / Fourier analysis
• frequency spectrum
  • about / Fourier analysis

G

• Gaussian pyramid / Using a bilateral filter for edge-aware smoothing
• generalization
  • about / The testing procedure
• graphical user interface (GUI) / Planning the app
• grayscale features / Grayscale features
• grayscale images
  • faces, detecting in / Detecting faces in grayscale images
• GTSRB dataset
  • about / The GTSRB dataset
  • URL / The GTSRB dataset
  • parsing / Parsing the dataset
• GUI base class
  • about / The GUI base class
  • GUI constructor / The GUI constructor
  • video streams, handling / Handling video streams
  • GUI layout / A basic GUI layout

H

• Haar-based cascade classifier
  • about / Haar-based cascade classifiers
• hand gesture recognition
  • about / Hand gesture recognition
  • convexity defects causes, distinguishing / Distinguishing between different causes of convexity defects
  • classifying, based on number of extended fingers / Classifying hand gestures based on the number of extended fingers
• hand gestures
  • tracking, in real time / Tracking hand gestures in real time
• hand region segmentation
  • about / Hand region segmentation
  • prominent depth of image center region, finding / Finding the most prominent depth of the image center region
  • morphological closing, applying / Applying morphological closing to smoothen the segmentation mask
  • connected components, finding / Finding connected components in a segmentation mask
• hand shape analysis
  • about / Hand shape analysis
  • contour of segmented hand region, determining / Determining the contour of the segmented hand region
  • convex hull of contour area, finding / Finding the convex hull of a contour area
  • convexity defects of convex hull, finding / Finding the convexity defects of a convex hull
• harmonics
  • about / Fourier analysis
• hidden layer
  • about / Deep architectures
• high-resolution images dataset
  • URL, for downloading / Setting up the app
• Histogram of Oriented Gradients (HOG)
  • about / Feature extraction, Histogram of Oriented Gradients
• HSV color space / Designing the warming/cooling effect
• hue
  • about / Color spaces
• hyperparameter exploration
  • about / Feature extraction

I

• image
  • cartoonizing / Cartoonizing an image
  • bilateral filter, used for edge-aware smoothing / Using a bilateral filter for edge-aware smoothing
  • prominent edges, detecting / Detecting and emphasizing prominent edges
  • prominent edges, emphasizing / Detecting and emphasizing prominent edges
  • colors, combining with outlines / Combining colors and outlines to produce a cartoon
• image plane
  • about / The pinhole camera model
• image rectification
  • about / Image rectification
• images pair
  • camera motion, estimating from / Estimating the camera motion from a pair of images
• information overflow
  • about / Visual saliency
• International Joint Conference on Neural Networks (IJCNN)
  • about / The GTSRB dataset
• intrinsic camera matrix
  • about / The pinhole camera model
• intrinsic camera parameters
  • estimating / Estimating the intrinsic camera parameters
  • camera calibration GUI / The camera calibration GUI
  • algorithm, initializing / Initializing the algorithm
  • image, collecting / Collecting image and object points
  • object points , collecting / Collecting image and object points
  • camera matrix, finding / Finding the camera matrix

K

• kernel trick
  • about / Support Vector Machine
• keypoints / Tasks performed by the app
• Kinect depth sensor
  • accessing / Accessing the Kinect 3D sensor
  • Kinect GUI / The Kinect GUI

L

• Laplacian / Detecting and emphasizing prominent edges
• learner
  • about / The training procedure
• learning rate
  • about / Deep architectures
• lookup tables
  • used, for implementing curve filter / Implementing a curve filter by using lookup tables
• loss function
  • about / Deep architectures
• Lukas-Kanade method
  • about / Point matching using optic flow

M

• mask / Creating a black-and-white pencil sketch
• matching procedure
  • result / Seeing the algorithm in action
• Mayavi
  • URL / 3D point cloud visualization
• mean-shift algorithm
  • objects, tracking with / Tracking objects with the mean-shift algorithm
• mean-shift tracking
  • about / Mean-shift tracking
  • bookkeeping, setting up for / Setting up the necessary bookkeeping for mean-shift tracking
• mean subtraction
  • about / Common preprocessing, Feature extraction
• median blur / Cartoonizing an image
• methods, Saliency class
  • Saliency.get_saliency_map / Planning the app
  • Saliency.get_proto_objects_map / Planning the app
  • Saliency.plot_power_density / Planning the app
  • Saliency.plot_power_spectrum / Planning the app
• MLP, for facial expression recognition
  • about / An MLP for facial expression recognition
• modules and scripts,app
  • gui / Planning the app
• modules and scripts, app
  • filters / Planning the app
  • filters.PencilSketch / Planning the app
  • filters.WarmingFilter / Planning the app
  • filters.CoolingFilter / Planning the app
  • filters.Cartoonizer / Planning the app
  • gui.BaseLayout / Planning the app, Planning the app
  • chapter1 / Planning the app
  • chapter1.FilterLayout / Planning the app
  • chapter1.main / Planning the app
  • feature_matching / Planning the app
  • feature_matching.FeatureMatching / Planning the app
  • gui / Planning the app
  • chapter3 / Planning the app
  • chapter3.FeatureMatchingLayout / Planning the app
  • chapter3.main / Planning the app
• modules and scripts, Kinect depth sensor
  • gestures / Planning the app
  • gestures.HandGestureRecognition / Planning the app
  • gui / Planning the app
  • gui.BaseLayout / Planning the app
  • chapter2 / Planning the app
  • chapter2.KinectLayout / Planning the app
  • chapter2.main / Planning the app
• Multi-Class classification
  • Support Vector Machine (SVM), using for / Using SVMs for Multi-class classification
• Multi-Layer Perceptrons (MLPs)
  • about / Multi-layer perceptrons
  • deep architectures / Deep architectures
  • training / Training the MLP
  • testing / Testing the MLP
  • script, running / Running the script
• MultiObjectTracker class
  • advance_frame method / Planning the app

N

• natural scene statistics
  • about / Natural scene statistics
• normalization
  • about / Common preprocessing, Feature extraction

O

• objects
  • tracking, with mean-shift algorithm / Tracking objects with the mean-shift algorithm
• one-vs-all strategy
  • about / Using SVMs for Multi-class classification
• one-vs-one strategy
  • about / Using SVMs for Multi-class classification
• OpenCV
  • dodging, implementing / Implementing dodging and burning in OpenCV
  • burning, implementing / Implementing dodging and burning in OpenCV
• OpenCV 3 / Creating a black-and-white pencil sketch
• opencv_contrib
  • URL / Tasks performed by the app
• optic flow
  • used, for point matching / Point matching using optic flow
• Oriented FAST and Rotated BRIEF (ORB) / Feature detection
• overfitting
  • about / The testing procedure, Deep architectures

P

• perceptron
  • about / Multi-layer perceptrons, The perceptron
• perspective transform
  • warping / Tasks performed by the app
• pinhole camera model
  • about / The pinhole camera model
  • reference link / The pinhole camera model
• players
  • tracking, automatically on soccer field / Automatically tracking all players on a soccer field
• Point Cloud Library
  • URL / 3D point cloud visualization
• point matching
  • performing, rich feature descriptors used / Point matching using rich feature descriptors
  • performing, optic flow used / Point matching using optic flow
• positive predictive value
  • about / Precision
• pre-trained cascade classifiers
  • about / Pre-trained cascade classifiers
  • using / Using a pre-trained cascade classifier
• precision
  • about / Precision
• prerequisites, app
  • main function / Planning the app
  • Saliency class / Planning the app
  • MultiObjectTracker class / Planning the app
• principal component analysis (PCA)
  • about / Common preprocessing, Feature extraction, Principal component analysis
• principal ray
  • about / The pinhole camera model
• process flow
  • about / The process flow
• proto-objects
  • detecting, in scene / Detecting proto-objects in a scene
  • bounding boxes, extracting for / Extracting bounding boxes for proto-objects
• pycaffe
  • reference link / Summary
• pylearn
  • reference link / Summary

R

• radially averaged power spectrum (RAPS)
  • about / Natural scene statistics
• random sample consensus (RANSAC) / Homography estimation
• ratio test / The ratio test for outlier removal
• read() method / Accessing the Kinect 3D sensor
• recall
  • about / Recall
• Region of Interest (ROI)
  • about / Parsing the dataset
• regression
  • about / Supervised learning
• regularization
  • about / The testing procedure
• rich feature descriptors
  • used, for point matching / Point matching using rich feature descriptors
• ridge operator / Detecting and emphasizing prominent edges

S

• Saliency class
  • methods / Planning the app
• Saliency map
  • generating, with spectral residual approach / Generating a Saliency map with the spectral residual approach
• saturation
  • about / Color spaces
• Scale-Invariant Feature Transform (SIFT) / Feature detection
• scale invariance
  • about / Natural scene statistics
• scale invariant / Feature detection
• scene
  • reconstructing / Reconstructing the scene
  • proto-objects, detecting in / Detecting proto-objects in a scene
• Scharr operator / Detecting and emphasizing prominent edges
• scikit-learn
  • reference link / Summary
• scikit-learn machine learning package
  • reference link / Testing the SVM
• score
  • about / The training procedure
• singular value decomposition (SVD) / Finding the camera matrices
• Sobel operator / Detecting and emphasizing prominent edges
• spatial domain
  • about / Fourier analysis
• spectral residual
  • about / Generating a Saliency map with the spectral residual approach
• spectral residual approach
  • Saliency map, generating with / Generating a Saliency map with the spectral residual approach
• Speeded-Up Robust Features (SURF)
  • about / Feature extraction, Speeded Up Robust Features
• stochastic gradient descent
  • about / Deep architectures
• strawlab, GitHub
  • reference link / 3D point cloud visualization
• supervised learning
  • about / Supervised learning
  • training procedure / The training procedure
  • testing procedure / The testing procedure
  • classifier base class / A classifier base class
• Support Vector Machine (SVM)
  • about / Support Vector Machine
  • using, for Multi-Class classification / Using SVMs for Multi-class classification
  • training / Training the SVM
  • testing / Testing the SVM
• support vectors
  • about / Support Vector Machine
• SURF
  • about / Tasks performed by the app, Feature detection
  • used, for detecting features in image / Detecting features in an image with SURF

T

• Theano
  • reference link / Summary
• Torch
  • reference link / Summary
• training set
  • about / The training procedure
• training set, facial expression recognition
  • assembling / Assembling a training set
  • screen capture, running / Running the screen capture
  • GUI constructor / The GUI constructor
  • GUI layout / The GUI layout
  • current frame, processing / Processing the current frame
  • training sample, adding / Adding a training sample to the training set
  • dumping, to file / Dumping the complete training set to a file
• triangulation
  • about / Reconstructing the scene
• true positives
  • about / Precision

U

• UC Irvine Machine Learning Repository
  • reference link / Summary
• underfitting
  • about / Deep architectures

V

• value
  • about / Color spaces
• Vapnik-Chervonenkis (VC dimension)
  • about / Deep architectures
• VisPy
  • URL / 3D point cloud visualization
• visual saliency
  • about / Visual saliency

W

• warming/cooling filter
  • generating / Generating a warming/cooling filter
  • color temperature / Generating a warming/cooling filter
  • color manipulation, via curve shifting / Color manipulation via curve shifting
  • curve filter, implementing with lookup tables / Implementing a curve filter by using lookup tables
  • warming/cooling effect, designing / Designing the warming/cooling effect
• wxPython 2.8
  • URL / Tasks performed by the app

Z

• zero-centering
  • about / Common preprocessing