Thuật ngữ AI
Từ điển đầy đủ về Trí tuệ nhân tạo
Gray-Level Co-occurrence Matrix (GLCM)
A statistical matrix that describes the frequency of occurrence of pixel pairs with specific intensity values at a given direction and distance, used to extract second-order textural features.
Haralick Texture Descriptors
A set of 14 statistical features calculated from the GLCM including contrast, homogeneity, energy, and correlation, used to quantify the textural properties of digital images.
Gabor Transform
A spatial-frequency analysis method using Gaussian filters modulated by sinusoids to capture textural features at different scales and orientations.
Local Binary Patterns (LBP)
A texture operator that encodes the micro-structural relationships between a central pixel and its neighbors into binary values, creating a robust histogram for texture classification.
Fractal Analysis
A mathematical approach measuring the complexity and self-similarity of textures at different scales, often used to quantify the roughness of natural surfaces.
Wavelets for Texture Analysis
A multi-resolution decomposition using wavelet functions to extract textural features at different spatial frequencies and orientations simultaneously.
Texture Invariant Moments
Mathematical descriptors based on the statistical moments of the intensity histogram, invariant to geometric transformations for robust texture recognition.
Spatial Autocorrelation
A statistical measure quantifying the correlation of pixel values at different spatial distances, characterizing the periodicity and regularity of textural patterns.
Texture Markov Models
Probabilistic approach modeling the statistical dependency between neighboring pixels to characterize and synthesize textures based on local contextual relationships.
2D Power Spectrum
Frequency representation obtained by Fourier transform revealing directional and periodic components of textures for their analysis and classification.
Laws' Texture Energy
Set of 5x5 convolution masks to detect fundamental textural primitives like levels, edges, spots, and ripples in images.
Texture Anisotropy
Measure of the directional dependency of textural features, quantifying how properties vary across different orientations in the image.
Texture Entropy
Informational measure quantifying the degree of disorder or unpredictability in the pixel intensity distribution of a textured region.
Convolutional Neural Networks for Texture
Deep learning architecture using automatically learnable filters to extract hierarchies of complex textural features from raw images.
Texture Segmentation
Process of partitioning an image into homogeneous regions based on similarities in textural features rather than solely on intensity differences.
Example-based Texture Synthesis
Technique generating new textures by sampling and recombining patches from a source texture while preserving local and global statistical characteristics.
Histogram of Oriented Gradients (HOG)
Descriptor counting occurrences of gradient orientation in localized portions of an image, effective for capturing directional texture structures.
3D Texture Analysis
Extension of texture analysis methods to the volumetric domain, considering three-dimensional relationships of voxels to characterize textured surfaces and volumes.