图像分割

分割图像

图像分割是将图像分割成多个部分或区域的过程。这种分割通常基于图像中像素的特征。例如，在图像中查找区域的一种方法是寻找像素值中的突变不连续性，这通常表示边缘。这些边缘可以定义区域。其他方法根据颜色值或纹理将图像分成多个区域。

App

色彩分割器	对彩色图像进行阈值化处理
图像分割器	通过细化区域来分割图像
三维体分割器	Segment 3-D grayscale or RGB volumetric images

函数

全部展开

使用分割一切模型 (SAM) 的分段

`imsegsam`	Perform automatic full image segmentation using Segment Anything Model (SAM) (自 R2024b 起)
`segmentAnythingModel`	Pretrained Segment Anything Model (SAM) for semantic segmentation (自 R2024a 起)

分割方法

`graythresh`	使用 Otsu 方法计算全局图像阈值
`multithresh`	Multilevel image thresholds using Otsu’s method
`otsuthresh`	使用 Otsu 方法的全局直方图阈值
`adaptthresh`	使用局部一阶统计量的自适应图像阈值
`grayconnected`	Select contiguous image region with similar gray values using flood-fill technique
`watershed`	分水岭变换
`activecontour`	使用主动轮廓 (snake) 区域增长方法将图像分割成前景和背景
`lazysnapping`	Segment image into foreground and background using graph-based segmentation
`grabcut`	Segment image into foreground and background using iterative graph-based segmentation
`imseggeodesic`	Segment image into two or three regions using geodesic distance-based color segmentation
`imsegfmm`	Binary image segmentation using fast marching method
`gradientweight`	Calculate weights for image pixels based on image gradient
`graydiffweight`	Calculate weights for image pixels based on grayscale intensity difference
`imsegkmeans`	基于 K 均值聚类的图像分割
`imsegkmeans3`	K-means clustering based volume segmentation
`imsegisodata`	ISODATA clustering based image segmentation (自 R2024b 起)
`superpixels`	图像的二维超像素过分割
`superpixels3`	3-D superpixel oversegmentation of 3-D image

显示分割结果

`imoverlay`	将二值掩膜刻入二维图像
`labeloverlay`	Overlay label matrix regions on 2-D image
`label2idx`	Convert label matrix to cell array of linear indices
`boundarymask`	Find region boundaries of segmentation

评估分割准确度

`jaccard`	Jaccard similarity coefficient for image segmentation
`dice`	Sørensen-Dice similarity coefficient for image segmentation
`bfscore`	Contour matching score for image segmentation

主题

快速入门

Get Started with Image Segmentation
Get started with tools for image segmentation, including Segment Anything Model, classical segmentation techniques, and deep learning-based semantic and instance segmentation.
Get Started with Segment Anything Model for Image Segmentation
Perform interactive image segmentation using Segment Anything Model (SAM) and deep learning.

使用图像分割器进行分割

Getting Started with Image Segmenter

Segment an image using different techniques, refine and save the binary mask, and export the segmentation code by using the Image Segmenter app.
Refine Segmentation Using Morphology in Image Segmenter
This example shows how to use the capabilities of the Image Segmenter app to polish the appearance of the mask image you created with the app.
Use Texture Filtering in Image Segmenter
This example shows how to use the Include Texture Features option in the Image Segmenter app to segment an image.

使用色彩分割器进行分割

Segment Image and Create Mask Using Color Thresholder
Segment an image based on color values and create a binary mask image using Color Thresholder.
Acquire Live Images in Color Thresholder
Perform color thresholding on an image acquired from a live USB webcam using the Color Thresholder app.

使用加博滤波器、超像素和其他技术进行分割

Plot Land Classification with Color Features and Superpixels
Classify land types based on color features using K-means clustering and superpixels.
Segment Lungs from 3-D Chest Scan
Perform a 3-D segmentation using active contours (snakes) and view the results using the Volume Viewer app.

使用三维体分割器进行分割

Create Binary Mask Using Volume Segmenter
This example shows how to segment a volume in the Volume Segmenter app.
Create Semantic Segmentation Using Volume Segmenter
This example shows how to create a semantic segmentation of a volume using the Volume Segmenter app.
Work with Blocked Images Using Volume Segmenter
Segment a volumetric image that could be too large to fit into memory by converting the volume to a blocked image.

精选示例

Segment Objects in Interactive ROI Using Segment Anything Model

Perform interactive segmentation of an object in a selected region of interest (ROI) of an image using the Segment Anything Model (SAM).

打开实时脚本

Texture Segmentation Using Gabor Filters

Use texture segmentation to identify regions based on their texture. The goal is to segment the dog from the bathroom floor. The segmentation is visually obvious because of the difference in texture between the regular, periodic pattern of the bathroom floor, and the regular, smooth texture of the dog's fur.