Index

A

• additive
  • about / A quick note on BGR
• application
  • modifying / Modifying the application
• artificial neural networks (ANN)
  • about / Artificial neural networks
  • statistical model / Artificial neural networks
  • perceptrons / Neurons and perceptrons
  • structure / The structure of an ANN
  • network layers, by example / Network layers by example
  • in OpenCV / ANNs in OpenCV
  • ANN-imal classification / ANN-imal classification
  • training epochs / Training epochs
  • digit recognition, handwritten / Handwritten digit recognition with ANNs
  • handwritten digit recognition / Handwritten digit recognition with ANNs
  • URL / Other parameters
  • improvements / Improvements
  • potential applications / Potential applications
• AT&T
  • URL / Performing face recognition

B

• background subtractors
  • about / Background subtractors – KNN, MOG2, and GMG
  • code / Back to the code
• Bag of words (BOW)
  • about / Bag-of-words, BOW in computer vision
  • in computer vision / BOW in computer vision
  • kmeans clustering / The k-means clustering
• BGR
  • about / A quick note on BGR
• Binary Robust Independent Elementary Features (BRIEF) algorithm / BRIEF
• Blue-green-red (BGR) / Reading/writing an image file
• BRIEF / Feature detection algorithms
• Brute-Force matching / Brute-Force matching

C

• calcBackProject function
  • about / The calcBackProject function
• calcHist function / The calcHist function
• Cameo
  • about / Project Cameo (face tracking and image manipulation), Cameo – an object-oriented design
  • video stream abstracting, managers.CaptureManager used / Abstracting a video stream with managers.CaptureManager
  • window abstracting, managers.WindowManager used / Abstracting a window and keyboard with managers.WindowManager
  • keyboard abstracting, managers.WindowManager used / Abstracting a window and keyboard with managers.WindowManager
  • class / Applying everything with cameo.Cameo
• camera frames
  • capturing / Capturing camera frames
  • displaying, in window / Displaying camera frames in a window
• CAMShift
  • about / Back to the code, CAMShift
• Canny
  • used, for edge detection / Edge detection with Canny
• car detection
  • about / Detecting cars, What did we just do?
  • sliding windows / SVM and sliding windows
  • SVM / SVM and sliding windows
  • in scene / Example – car detection in a scene
  • testing / Dude, where's my car?
• car detection, in scene
  • detector.py, examining / Examining detector.py
  • training data, associating with classes / Associating training data with classes
• cascade / Conceptualizing Haar cascades
• channels
  • depth map / Capturing frames from a depth camera
  • point cloud map / Capturing frames from a depth camera
  • disparity map / Capturing frames from a depth camera
  • valid depth mask / Capturing frames from a depth camera
• circle detection
  • about / Line and circle detection, Circle detection
• clustering / The k-means clustering
• color histograms
  • about / Color histograms
  • calcHist / The calcHist function
  • calcBackProject / The calcBackProject function
  • in summary / In summary
• color spaces
  • converting, between / Converting between different color spaces
• comma-separated value (CSV) file / Preparing the training data
• contours
  • detection / Contour detection
  • bounding box / Contours – bounding box, minimum area rectangle, and minimum enclosing circle
  • minimum enclosing circle / Contours – bounding box, minimum area rectangle, and minimum enclosing circle
  • minimum area rectangle / Contours – bounding box, minimum area rectangle, and minimum enclosing circle
  • convex contours / Contours – convex contours and the Douglas-Peucker algorithm
  • douglas-peucker algorithm / Contours – convex contours and the Douglas-Peucker algorithm
• Contrib modules
  • installing / Installing the Contrib modules
• convergence
  • URL / Meanshift and CAMShift
• convex contours / Contours – convex contours and the Douglas-Peucker algorithm
• convolution matrix
  • about / Custom kernels – getting convoluted
• copy operation
  • masking / Masking a copy operation

D

• depth camera
  • modules, creating / Creating modules
  • frames, capturing / Capturing frames from a depth camera
• depth estimation
  • with normal camera / Depth estimation with a normal camera
• depth map
  • about / Capturing frames from a depth camera
• Difference of Gaussians (DoG) / Feature extraction and description using DoG and SIFT
• Discrete Fourier Transform (DFT) / The Fourier Transform
• disparity map / Capturing frames from a depth camera
  • mask, creating from / Creating a mask from a disparity map
• douglas-peucker algorithm / Contours – convex contours and the Douglas-Peucker algorithm

E

• edge
  • detecting / Edge detection
  • detection, with Canny / Edge detection with Canny
• Eigenfaces
  • about / Recognizing faces
  • URL / Recognizing faces
  • recognition, performing / Performing an Eigenfaces recognition
• Extended Yale or Yale B
  • URL / Performing face recognition

F

• face detection
  • performing, OpneCV used / Using OpenCV to perform face detection
  • performing, on still image / Performing face detection on a still image
  • performing, on video / Performing face detection on a video
  • performing / Performing face recognition
  • data, generating / Generating the data for face recognition
  • faces, recognizing / Recognizing faces, Loading the data and recognizing faces
  • training data, preparing / Preparing the training data
  • data, loading / Loading the data and recognizing faces
  • Eigenfaces recognition, performing / Performing an Eigenfaces recognition
  • face recognition, performing with Fisherfaces / Performing face recognition with Fisherfaces
  • face recognition, performing with LBPH / Performing face recognition with LBPH
  • results, discarding with confidence score / Discarding results with confidence score
• FAST / Feature detection algorithms
• Fast Fourier Transform (FFT) / The Fourier Transform
• Fast Library for Approximate Nearest Neighbors (FLANN) / FLANN-based matching
  • matching, with homography / FLANN matching with homography
• feature detection algorithms
  • about / Feature detection algorithms
  • features, defining / Defining features
  • corners / Detecting features – corners
  • DoG used / Feature extraction and description using DoG and SIFT
  • SIFT used / Feature extraction and description using DoG and SIFT
  • keypoint, anatomy / Anatomy of a keypoint
  • Fast Hessian used / Feature extraction and detection using Fast Hessian and SURF
  • SURF used / Feature extraction and detection using Fast Hessian and SURF
  • ORB feature detection / ORB feature detection and feature matching
  • Features from Accelerated Segment Test (FAST) algorithm / FAST
  • Binary Robust Independent Elementary Features (BRIEF) algorithm / BRIEF
  • Brute-Force matching / Brute-Force matching
  • feature matching, with ORB / Feature matching with ORB
  • K-Nearest Neighbors matching, using / Using K-Nearest Neighbors matching
  • Fast Library for Approximate Nearest Neighbors (FLANN) / FLANN-based matching
  • FLANN matching, with homography / FLANN matching with homography
• features / Conceptualizing Haar cascades
• Features from Accelerated Segment Test (FAST) algorithm / FAST
• Fisherfaces
  • about / Recognizing faces
  • URL / Recognizing faces
  • face recognition, performing with / Performing face recognition with Fisherfaces
• Fourier Transform
  • about / The Fourier Transform
  • high pass filter (HPF) / High pass filter
  • low pass filter (LPF) / Low pass filter
• frames
  • capturing, from depth camera / Capturing frames from a depth camera
• functional programming (FP) solutions / A brief digression – functional versus object-oriented programming

G

• geometric multigrid (GMG) / Background subtractors – KNN, MOG2, and GMG
• GrabCut algorithm
  • used, for object segmentation / Object segmentation using the Watershed and GrabCut algorithms
  • foreground detection, example / Example of foreground detection with GrabCut

H

• Haar cascades
  • conceptualizing / Conceptualizing Haar cascades
  • data, getting / Getting Haar cascade data
• handwritten digit recognition
  • with ANNs / Handwritten digit recognition with ANNs
  • MNIST database / MNIST – the handwritten digit database
  • training data, customized / Customized training data
  • initial parameters / The initial parameters
  • training epochs / Training epochs
  • other parameters / Other parameters
  • mini-libraries / Mini-libraries
  • main file / The main file
• Harris / Feature detection algorithms
• hidden layer
  • about / The hidden layer, The hidden layer
• high pass filter (HPF) / High pass filter
• histogram back projection / The calcBackProject function
• Histogram of Oriented Gradients (HOG)
  • about / HOG descriptors
  • scale issue / The scale issue
  • location issue / The location issue
  • image pyramid / Image pyramid
  • sliding windows / Sliding windows
  • nonmaximum (or nonmaxima) suppression / Non-maximum (or non-maxima) suppression
  • support vector machines / Support vector machines
• Hue, Saturation, Value (HSV)
  • about / Converting between different color spaces

I

• I/O scripts
  • basic / Basic I/O scripts
  • image file, reading / Reading/writing an image file
  • image file, writing / Reading/writing an image file
  • image and raw bytes, conversions / Converting between an image and raw bytes
  • image data, accessing with numpy.array / Accessing image data with numpy.array
  • video file, writing / Reading/writing a video file
  • video file, reading / Reading/writing a video file
  • camera frames, capturing / Capturing camera frames
  • images, displaying in window / Displaying images in a window
  • camera frames, displaying in window / Displaying camera frames in a window
• image
  • about / Converting between an image and raw bytes
  • displaying, in window / Displaying images in a window
• image data
  • accessing, numpy.array used / Accessing image data with numpy.array
• image descriptors
  • saving, to file / Saving image descriptors to file
  • matches, scanning for / Scanning for matches
• image file
  • reading / Reading/writing an image file
  • writing / Reading/writing an image file
• image segmentation
  • with Watershed algorithm / Image segmentation with the Watershed algorithm
• input layer / The input layer, The input layer
• in summary
  • about / In summary

K

• K-Nearest Neighbors (KNN) / Using K-Nearest Neighbors matching, Background subtractors – KNN, MOG2, and GMG
• Kalman filter
  • about / The Kalman filter, Where do we go from here?
  • predict phase / Predict and update
  • update phase / Predict and update
  • example / An example
  • real-life example / A real-life example – tracking pedestrians
  • Pedestrian class / The Pedestrian class
  • main program / The main program
• Kalman filter, real-life example
  • about / A real-life example – tracking pedestrians
  • application work flow / The application workflow
  • functional, versus object-oriented programming / A brief digression – functional versus object-oriented programming
• kernels
  • about / High pass filter
  • custom kernels / Custom kernels – getting convoluted
• keyboard
  • abstracting, managers.WindowManager used / Abstracting a window and keyboard with managers.WindowManager
• kmeans clustering / The k-means clustering

L

• Lasagna / Other parameters
• learning algorithms
  • about / The learning algorithms
  • supervised learning / The learning algorithms
  • unsupervised learning / The learning algorithms
  • reinforcement learning / The learning algorithms
• lines detection
  • about / Line and circle detection, Line detection
• Local Binary Pattern
  • URL / Recognizing faces
• Local Binary Pattern Histograms
  • face recognition, performing with / Performing face recognition with LBPH
• Local Binary Pattern Histograms (LBPH)
  • about / Recognizing faces
• low pass filter (LPF) / Low pass filter

M

• magnitude spectrum / The Fourier Transform
• mask
  • creating, from disparity map / Creating a mask from a disparity map
• Meanshift
  • about / Meanshift and CAMShift
• mixture of Gaussians (MOG2) / Background subtractors – KNN, MOG2, and GMG
• MNIST
  • about / MNIST – the handwritten digit database
  • URL / MNIST – the handwritten digit database
• modules
  • creating / Creating modules
• multilayer perceptron / ANNs in OpenCV

N

• network layers
  • by example / Network layers by example
  • input layer / The input layer
  • output layer / The output layer
  • hidden layer / The hidden layer
• nonmaximum (or nonmaxima) suppression / Non-maximum (or non-maxima) suppression
• nonmaximum suppression (NMS) / Edge detection with Canny
• numpy.array
  • used, for accessing image data / Accessing image data with numpy.array

O

• object detection
  • about / Object detection and recognition techniques
• object detection, techniques
  • Histogram of Oriented Gradients (HOG) / HOG descriptors
  • people detection / People detection
  • creating / Creating and training an object detector
  • training / Creating and training an object detector
  • Bag of words (BOW) / Bag-of-words
• objects
  • moving objects, detecting / Detecting moving objects
  • motion detection / Basic motion detection
• object segmentation
  • Watershed algorithm used / Object segmentation using the Watershed and GrabCut algorithms
  • GrabCut algorithm used / Object segmentation using the Watershed and GrabCut algorithms
• Open Source Computer Vision (OCV)
  • building, from source / Building OpenCV from a source
  • documentation, URL / Finding documentation, help, and updates
  • community, URL / Finding documentation, help, and updates
• Oriented FAST and Rotated BRIEF (ORB) / Feature detection algorithms
  • feature detection and feature matching / ORB feature detection and feature matching
• OS X
  • installation on / Installing on OS X
  • MacPorts, using with ready-made packages / Using MacPorts with ready-made packages
  • MacPorts, using with own custom packages / Using MacPorts with your own custom packages
  • Homebrew, using with ready-made packages / Using Homebrew with ready-made packages (no support for depth cameras)
  • Homebrew, using with own custom packages / Using Homebrew with your own custom packages
• output layer / The output layer, The output layer
• overfitting / The hidden layer

P

• Pedestrian class
  • about / The Pedestrian class
• perceptrons
  • about / Neurons and perceptrons
• point cloud map / Capturing frames from a depth camera
• Principal Component Analysis (PCA)
  • URL / Recognizing faces
• PyBrain / Other parameters
• pyramid / Image pyramid

R

• raw bytes
  • about / Converting between an image and raw bytes
• red-green-blue (RGB) / Reading/writing an image file
• regional handwriting variations
  • URL / Improvements
• region of interest (ROI) / The Pedestrian class
• regions of interests (ROI) / Accessing image data with numpy.array
• reinforcement learning / The learning algorithms
• resilient back propagation (RPROP) / Other parameters
• ridge / Defining features

S

• samples
  • running / Running samples
• Scale-Invariant Feature Transform (SIFT) / Feature extraction and description using DoG and SIFT
• semiglobal block matching / Depth estimation with a normal camera
• setup tools
  • selecting / Choosing and using the right setup tools
  • installation, on Windows / Installation on Windows
  • installation, on OS X / Installing on OS X
  • installation, on Ubuntu and derivatives / Installation on Ubuntu and its derivatives
  • installation, on Unix-like systems / Installation on other Unix-like systems
• shapes
  • detecting / Detecting shapes
• SIFT / Feature detection algorithms
• sigmoid function / Neurons and perceptrons
• sliding windows / Sliding windows
• Stanford University
  • URL / Detecting cars
• statistical model / Artificial neural networks, ANNs in OpenCV
• StereoSGBM parameters
  • minDisparity / Depth estimation with a normal camera
  • numDisparities / Depth estimation with a normal camera
  • windowSize / Depth estimation with a normal camera
  • P1 / Depth estimation with a normal camera
  • P2 / Depth estimation with a normal camera
  • disp12MaxDiff / Depth estimation with a normal camera
  • preFilterCap / Depth estimation with a normal camera
  • uniquenessRatio / Depth estimation with a normal camera
  • speckleWindowSize / Depth estimation with a normal camera
  • speckleRange / Depth estimation with a normal camera
• subtractive color model
  • about / A quick note on BGR
• supervised learning / The learning algorithms
• support vector machines (SVM) / Non-maximum (or non-maxima) suppression
  • URL / Support vector machines
• SURF / Feature detection algorithms

U

• Ubuntu
  • repository, using / Using the Ubuntu repository (no support for depth cameras)
• University of Illinois
  • URL / Detecting cars
• Unix-like systems
  • installation on / Installation on other Unix-like systems
• unsupervised learning / The learning algorithms

V

• valid depth mask / Capturing frames from a depth camera
• Variante Ascari / Feature extraction and description using DoG and SIFT
• video file
  • reading / Reading/writing a video file
  • writing / Reading/writing a video file
• video stream
  • abstracting, managers.CaptureManager used / Abstracting a video stream with managers.CaptureManager

W

• Watershed algorithm
  • used, for object segmentation / Object segmentation using the Watershed and GrabCut algorithms
  • used, for image segmentation / Image segmentation with the Watershed algorithm
• window
  • abstracting, managers.WindowManager used / Abstracting a window and keyboard with managers.WindowManager
• Windows
  • installation on / Installation on Windows
  • binary installers, using / Using binary installers (no support for depth cameras)
  • CMake , using / Using CMake and compilers
  • compilers, using / Using CMake and compilers
• window size / Conceptualizing Haar cascades

Y

• Yale face database (Yalefaces)
  • URL / Performing face recognition