Digital image processing vocabulary

Digital image processing vocabulary

Algebraic operation algebra is an image processing operation that includes the sum, difference, product, and quotient of the two images corresponding to pixels.

Aliasing is a kind of artificial trace generated when the pixel distance between the image and the image details are too large.

A part of an arc chart, indicating a collection of connected pixels in a curve segment.

Binary Image Binary Images only have two levels of gray digital images (usually 0 and 1, black and white)

Blur blur reduces the image definition caused by defocus, low-pass filtering, and camera motion.

Border borders the first, last, or column of an image.

Boundary Chain Code boundary chaincodes define the direction sequence of an object boundary.

Boundary pixel boundary pixel at least corresponds to an internal pixel adjacent to a background pixel (comparison: External pixel, internal pixel)

Boundary Tracking boundary tracking is an image segmentation technique that explores the arc from a pixel order to the next pixel and detects the arc.

Brightness indicates the amount of light emitted or emitted from an object in the image.

Change Detection change detection compares the pixels of two quasi-images through subtraction to detect the differences between objects.

Class class see module or Class

The closed curve is a curve in the same position at the first and last points.

Cluster clustering, a set of points that the cluster is close to in the space (such as in the feature space.

Cluster analysis cluster analysis checks, measures, and descriptions clustering in space.

Concave concave objects are concave objects that contain at least two points in an object, and their links cannot be completely contained in the object (the antonym is convex)

Connected

Contour encoding contour encoding is an image compression technology that only encodes its boundary for areas with even gray scale.

The differences between the average brightness (or gray scale) of a contrast object and its surrounding background

Contrast stretch contrast extension a linear grayscale Transformation

A convex object is a convex object. A straight line connecting any two points inside the object falls inside the object.

Convolution convolution is an operation that combines two functions into a third function. convolution depicts the operation of a linear shift-free system.

Convolution volution kernel convolution kernel 1, a two-dimensional digital array used for digital image convolution filtering. 2, a function for Convolution with images or signals.

Curve 1, a continuous path of space, 2 represents a set of pixels of a path (see the arc, closed curve ).

Deblurring deblur: an operation to reduce image blur and sharpen image details. 2. Removing or downgrading the blur of an image is usually a step of image restoration or reconstruction.

Demo-rule decision rules are used in pattern recognition to assign objects in images to a certain number of rules or algorithms. This assignment is based on object feature measurements.

Digital Image 1 represents an integer array of scene images. 2 is a two-dimensional or higher-dimensional sampling and quantization function, which is produced by continuous images of the same dimension, 3. Samples a continuous function on a rectangle (or other) network and quantifies the value at the sampling point.

Digital image processing digital image processing: digital processing of images; operations on image information by computers.

Digitization digitization converts a scene image into a digital form.

Edge edge 1 is a region where gray scale changes in the image, and 2 is a pixel set on an arc. pixels on the other side are significantly different from those on the other side.

Edge Detection an image segmentation technique that identifies edge pixels by checking the neighborhood.

Edge enhancement an image processing technique that enhances the image edge by comparing and expanding the pixels on both sides of the edge.

Edge Image each pixel in an edge image is either marked as an edge or non-edge.

Edge linking is an image processing technology that connects edge pixels into edges in edge images.

Edge Operator Edge operators represent edge pixels in an image.

Edge pixel: edge pixels on the edge

Enhance Enhances contrast or subjective visibility.

Exterior pixels external pixels in a binary image, pixels outside the object (relative to internal pixels)

False Negative negative false recognition: in two types of pattern recognition, labeling an object as a false classification of a non-Object

False Positive positive and false positive recognition: in two types of pattern recognition, labeling an object that is not part of an object as an incorrect classification of the object.

A feature of a feature object that can be measured. Helps to classify objects. Such as size, texture, and shape.

Feature Extraction: a step in Feature Detection Pattern Recognition. In this step, the related measurement of the object is calculated.

Feature Selection feature selection is a step in pattern recognition system development. It is intended to study whether quality or observation can be used to assign objects to a certain category.

Feature Space see measurement space

Fourier transform Fourier transformation uses the complex exponent as a linear transformation of the kernel function.

Geometric correction ry correction uses a geometric transformation to eliminate geometric distortion of an image restoration technology.

The gray level Level 1 is connected to the pixels of a digital image, which indicates the brightness of the original scene point of the pixel. 2. Digital Measurement of the local properties of an image at a pixel position.

Gray Scale: a set of all possible gray levels in a digital image

A function of gray-scale transformation in vertex operations. It establishes the relationship between the input gray scale and the corresponding output gray scale.

Hankel transform Hankel transformation

The harmonic signal has a complex signal composed of a cosine real part and a sine imaginary part of the same frequency.

Hermite function the Hermite function has a complex function with an even real and odd virtual part.

Highpass filtering high-pass filtering image enhancement (usually convolution) operation, which improves the high-frequency part compared to the low-frequency part.

Hole holes in a binary image are connected background points completely surrounded by the object's Interior Point

The unified representation of an image to a physical scene or other images is called an image.

Image Compression eliminates image redundancy or any process similar to an image, which is used to represent images in a more compact form

Image Coding Image Encoding converts an image into another recoverable form (such as compression)

Image Enhancement image enhancement is designed to improve the visual appearance of an image.

Image matching: The process of quantifying and comparing the similarity between the two images.

Image-Processing operation a series of steps for converting an input image into an output image

Image reconstruction process of constructing or recovering an image from a non-image form

Image Registration image accurately performs geometric operations on one image and another image of the same image to align the object.
Image Restoration the process of restoring an image to its original state through the process of inverse image degradation.

Image segmentation image segmentation 1. Detect and outline the processing of objects of interest in the image; 2. Divide the image into unconnected areas. Generally, these regions correspond to objects and the background of objects.

The pixels in interior pixel are in the interior of an object in a binary image (relative to the boundary pixel, external pixel)

Interpolation interpolation determines the sampling function between sampling points.

Kernel Core

Line Detection: an image segmentation technique that identifies linear pixels by checking the adjacent area

Line pixel linear pixels are located on an arc close to a straight line.

Local operation: the local operation is based on the pixel gray scale of a neighboring area of the input pixel, which determines the image processing operation of the output gray scale of the pixel. It is similar to the calculation of the neighboring area (comparison: vertex operation)

Local property: the characteristics of interest (such as the brightness or color of an optical image; non-Optical Image Height, temperature, density, etc.) that change along with position in the image)

Lossless Image Compression distortion-free image compression allows you to completely reconstruct any image compression technology of the original image.

Lossy Image Compression distortion image compression due to inclusion approximation, cannot accurately reconstruct any image compression technology of the original image

Matched Filtering matching and filtering the existence and location of a specific object in the image is detected by a matching wave.

Measurement space: in pattern recognition, the measurement space contains the n-dimensional vector space of all possible measurement vectors.

Misclassification in pattern recognition, mistaken object category

Multispectral Image multi-spectral images a set of images of the same scene, each of which is produced by the different band radiation of the electromagnetic spectrum

A set of neighborhood neighboring pixels

Neighoorhood operation

Noise noise is an irrelevant part of the image that hinders the identification and interpretation of interest data.

Noise noise reduction any processing of noise in a pair of images

An object is a set of connected pixels in a two-value image in pattern recognition. It usually corresponds to an object in the image.

Optical Image results in the projection of light from a scene to a surface through an optical device such as a lens
Pattern indicates the common and meaningful regularity of a class member, which can be measured and used to classify objects of interest.

The pattern class can assign an object any class of an incompatible pre-defined class set.

Pattern classification process of assigning objects to pattern classes

Pattern recognition pattern recognition: automatic or semi-automatic detection, measurement, and classification of objects in an image

Abbreviation of Pel pixel image Element

The peripheral distance of the perimeter around the boundary of an object

Picture element: the minimum unit of a pixel digital image. It is the basic unit of a digital image.

Abbreviation of pixel image Element

The point operation only determines the image processing operation of the output gray value of the pixel based on the input gray value of the corresponding pixel (comparison: neighborhood operation)

Quantitative image analysis image quantitative analysis the process of extracting quantitative data from a digital image

Quantization quantifies the process of assigning the partial characteristics of an image to elements in a gray set at each pixel.

Connected subset in a region Image

Region Growing Area Growth an image segmentation technique that repeatedly seeks the Union of adjacent sub-areas with similar gray or textures to form Areas

Registered match 1: the state of the two or more images have been visually tuned, and the objects in the two images are consistent.

Registered image: The two (or more) images of the same scene in the same image have the same position as each other, so that the object has the same image position.

Resolution resolution 1: Minimum separation distance between the points that can be distinguished in optics; 2. in image processing, it refers to the degree to which adjacent point objects in the image can be distinguished.

Run stroke in image encoding, connected pixel sequences with the same gray scale

Run Length: the number of pixels in the itinerary.

Run Length Encoding stroke-encoded image lines are represented by travel sequences. Each stroke is defined by a given travel length and grayscale value.

Sampling sampling (based on the sampling network) divides an image into pixels and measures its local characteristics (such as brightness and color ).

A distinctive layout of objective objects in scene scenes

Sharp: easy resolution of image details

Sharpening is an image processing technology used to enhance image details.

Sigmoid Function S function is a type of function in the form of S. It is a gray-scale change function. It can also be used in processing unit (PE) functions in the Neuron Network.

Sinusoidal sine type function type with sine function shape

A graphic processing technology that uses smoothing to smoothly reduce the image's details. It is usually used for noise reduction.

Statistical Pattern Recognition uses probability and statistical methods to assign an object to pattern recognition.

Structural Pattern Recognition Structure Pattern recognition is a pattern recognition method for describing and classifying objects and expressing objects as basic elements and their relationships

Syntactic Pattern Recognition: A structure pattern recognition method that uses natural or artificial language patterns to define elements and their relationships

System system receives input and outputs

In image processing, texture represents an attribute of the spatial organization of the Gray Scale Amplitude and its local changes in the image.

A two-value image processing technique for thinning to refine the fine curve of an object to a (single pixel width)

The threshold is used to produce a specific gray scale of a binary image.

Thresholding: the process of generating a binary image with a gray scale image. If the gray scale of the input pixel is greater than the given threshold, the output pixel is assigned 1; otherwise, the value is 0.

The transter function is a frequency function that expresses the amplitude ratio of the sine input signal at each frequency in a linear shift-free system and transmits it to the output signal.