OpenCV: Master OpenCV 3 Application Development Using Python

This video gives an overview of the entire course.

Before we jump into OpenCV functionalities, we need to understand why those functions were built. Let’s understand how the human visual system works so that we can develop the right algorithms.

•         Learn how the human eye captures information
•         See why machines find it difficult to understand image content

 Real-life problems require us to use many blocks together to achieve the desired result. So, we need to know what modules and functions to use. Let's understand what OpenCV can do out of the box.

•         Take a look at tasks we can do with OpenCV

Now that we know what tasks we can do with OpenCV, Let’s see how to get OpenCVupandrunning on various operating systems, viz. Windows, Mac and Linux.

•         Install OpenCV on Windows
•         Install OpenCV on Mac OS X
•         Install OpenCV on Linux

We are going to use CMake to configure and check all the required dependencies of our project. So, let’s learn basic CMake configuration files and creating a library.

•         Take a look CMakelists.txt file
•         Use add_library function

CMake has the ability to search our dependencies and external libraries, giving us the facility to build complex projects depending on external components in our projects and by adding some requirements. One of the most important dependency is, of course, OpenCV. Let’s learn how to add it to our projects.

•         Add OpenCVdependencyto all projects

 Now that we know managing dependencies, let’s take a look at a bit more complex script. This video we will show us a more complex script that includes subfolders, libraries, and executables, all in only two files and a few lines, as shown in the script.

•         Define two CMakelists.txt files
•         Write a new library in the CMakeLists.txt file in the UTILS folder
•         Use the add_definitions(-DLOG) function

The most important structure in computer vision is without any doubt the images. The image in computer vision is a representation of the physical world captured with a digital device. Let’s now learn about images and matrices.

•         Learn about pixels
•         Take a look at matrix format and matrix storage

After the introduction of matrix, we are ready to start with the basics of the OpenCV code. This video will guide us how to read and write images.

•         Take a look at imread and IM imwrite function
•         Use .cols and .rows attributes of a matrix
•         Useimshow function to show the image

We now know how to read and write images but reading video can be a bit tricky. This video introduces us reading a video and camera with simple example.

•         Take a look at the Command Line Parser class
•         Use the .get function
•         Show the frames with the named Window function

We have learned about the Mat and Vec3b classes, but we need to learn about otherclasses as well. In this video, we will learn about the most basic object types required in most of the projects.

•         Take a look at different object types

In many applications, such as calibration or machine learning, when we are done with the calculations, we need to save the results in order toretrieve them in the next executions. Before we finish this section, we will explore the OpenCV functions to storeand read our data.

•         Use mat constructor
•         Use zeros, ones, and eye function
•         Write to a file storage

 OpenCV has its own cross-operating system user interface that allows developers to create their own applications without the need to learn complex libraries for theuser interface. This video will introduce the OpenCV user interface and help us creating a basic UI with OpenCV.

•         Learn about the OpenCV user interface
•         Create a basic user interface with OpenCV

 The QT user interface gives more control and options to work with images. Let’s explore the interface and learn how to use it.

•         Take a look at toolbar buttons
•         Use displayOverlay function
•         Learn about status bar

Mouse events and slider controls are very useful in Computer Vision and OpenCV. Using these controls, users can interact directly with the interface and change the properties of their input images or variables. However, using these controls can be a bit tricky. Let’s see how to use them.

•         Create a variable to save the slider position
•         Load an imageand create a new window calledLena
•         Create the slider

Now that we have learned how to create normal or QT interfaces and interact with them using a mouse and slider, let’s see how we can create different types of buttons to add more interactivity.

•         Create a basic sample with buttons
•         Apply blur filter, a sobel filter, and a color conversion to gray

OpenCV includes OpenGL support which is a graphical library that is integrated in graphic cards as a standard. OpenGL allows us to draw from 2D to complex 3D scenes. This video shows us how to use OpenGL support.

•         Create a window with OpenGL support
•         Draw a rotated plane that shows the web camera frames

Prepare a CMake script file that enables us to compile our project, structure, and executable.

•         Generate CMakeLists.txt file
•         Finad the OpenCV library and show a message about the OpenCV library version
•         Add the source files to link them to OpenCV library.

The main graphical user interface can be used in the application to create single buttons.

•         Add the required OpenCV headers
•         Set up the instructions and print the help message
•         Show the input image and wait for a key press to finish our application

Histogram is a statistical graphic representation of variable distribution that allows us to understand the density estimation and probability distribution of data.

•         Create three matrices to process each input image channel
•         Calculate the histogram using the OpenCVcalcHist function
•         Show the histogram image with the imshow function

Image equalization obtains a histogram with a uniform distribution of values.

•         Convert the input BGR image into YCrCb using the cvtColor function
•         Split the YCrCb image into different channels matrices
•         Merge the resulted channels and convert the result to the BGR format

Lomography is a photographic effect used in different mobile applications, such as Google Camera or Instagram.

•         Manipulate the red color with a curve transform
•         Split our input image by channels using the split function
•         Convert the flat image matrix result to an 8-bit image

The Cartoonize effect creates an image that looks like a cartoon

•         Detect the most important edges of the image
•         Multiply our edges' result image by the color image
•         Merge the color and edges' results

Isolating different parts or objects in a scene.

•         Preprocess The image
•         Segment the image

Create our new application.

•         Enable user selection using command line parser class
•         Checks the parameters
•         Check whether the input image is correctly loaded

Extract the information from image.

•         Remove noise
•         Remove the background using the light pattern
•         Binarize the image

Extract each region of interest of our image where our target objects appear.

•         Apply each algorithm to the next binarized image
•         Return an integer with the number of detected labels
•         Draw each detected contour that we detect

Pattern recognition and the learning theory in artificial intelligence and are related to computational statistics.

•         Use a supervisedlearning classification algorithm
•         Learn and Categorize results obtain from machine learning

 We will learn how to implement our own application that uses machine learning to classify objectsin a slide tape.

•         See the basic structure of Computer Vision applications with machine learning
•         Preprocess involves light removal conditions, noise, thresholding, blur and so on
•         Extract the regions of image and isolate each one as a unique object

We will be able to recognize different objects to send notifications to a robot or put each one in different boxes.

•         Preprocessand Segment the image
•         Create an SVM model and train our SVM model with the training feature vector.
•         For each object detected: Extract the features and predict with an SVM model

 To extract the features of each object

•         We will create a function that has one image as the input and returns two vectors of left and top position for each object detected in the image as parameters

•         We need to create the output vector variable and contours variable that are to be used in our FindContoursalgorithm segmentation

•         Show the detected objects in a window for the user feedback

It is simply a concatenation of a set of weak classifiers that can be used to create a strong classifier.

•         Extract these features
•         Pass them through a cascade of classifiers
•         Check all the different rectangular sub regions and keep discarding the ones that don't have faces in them

You have to avoid huge redundancy during the area computation, to avoid this, we can use integral images

•         Pick up the right values
•         Substitute them in the previous equation
•         Extract the features

You have to load the cascade file and use it to detect the faces in an image.

•         Read input frames from the webcam
•         Convert image to grayscale
•         Extract the region of interest to overlay the mask

 You have to overlay sunglass on face.

•         Run the eye detector
•         Adjust, the size of the sunglasses
•         Overlay the sunglasses mask

You have to track nose, mouth and ears.

•         Use a nose detector
•         Overlay a nose
•         Cascade files used to track different parts

The background subtraction technique performs really well where we need to detect moving objects in a static scene.

•         Build a model of the background
•         Background model is then used to detect background pixels
•         The difference between background model and image is computed

We cannot keep a static background image that can be used to detectobjects.

•         The difference between successive frames and displaying the difference
•         Take the frame differences and apply a bitwise AND operator
•         Resize the frame and convert it to grayscale

Formulating and implementing a mixture of gaussians.

•         Identify whether the data has multiple groups inside the subpopulation
•         We represent each subpopulation using the Gaussian function

 Morphological Image processing is used in processing the shapes of features in the image.

•         Use a structuring element to modify an image
•         Use erosion operation
•         Use dilation operation

 To apply various morphological operators on image.

•         Apply erosion followed by dilation
•         use the morphological gradient
•         Apply Top-Hat transform and Black Hat transform

Understand what characteristics can be used to make our tracking robust and accurate.

•         Use colorspaces to come up with a good visual tracker
•         Take the pixels of an image to the HSV space
•         Use colorspace distances and threshold in this space thresholding to track a given object

We want to randomly pick an object, learn the characteristics of the selected object and track it automatically.

•         Select a bunch of points based on the color histogram
•         We move the bounding box of the object to a new location so that the new centroid becomes the center
•         Use CAMShift algorithm

Detect interest points in the image.

•         Construct a 2 x 2 matrix based on partial derivatives
•         Check whether this patch has a corner in it
•         Compute the intensity difference between our patch and all those neighboring patches

Improve the overall quality of image

•         Find the N strongest corners in the given image
•         Use goodFeaturesToTrackto detect the corners
•         Draw circles on these points to display the output image

Tracking individual feature points across successive frames in the video.

•         Choose a bunch of feature points
•         Track them through the video stream
•         Compute the displacement vectors and show the motion of those key points

Basics of OCR.

•         Text preprocessing and segmentation
•         Text identification

Classification results can be improved greatly if the input text is clear so Adjust the text.

•         Threshold the image
•         Text segmentation
•         Text extraction and skew adjustment

Install Tesseract on Windows or Mac.

•         Download the installer
•         Choose a suitable location for your installation
•         set up Tesseract dependencies

Studying tesserct API.

•         Create a OCR function
•         Call the SetImagemethod with the signature
•         Send the output to a file