Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Multi-scale image segmentation method based on hierarchical region merging

An image segmentation, multi-scale technology, applied in image analysis, image data processing, instruments, etc., can solve the problems of computational complexity and large amount of calculation, and achieve the effect of improving objectivity

Active Publication Date: 2019-11-29
SOUTHWEST JIAOTONG UNIV
View PDF6 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although better results have been achieved, the machine learning method requires more features, and the computational complexity and amount of calculation generated during the training process are large

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-scale image segmentation method based on hierarchical region merging
  • Multi-scale image segmentation method based on hierarchical region merging
  • Multi-scale image segmentation method based on hierarchical region merging

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] refer to figure 1 , a multi-scale image segmentation method based on hierarchical region merging, using existing multi-scale image segmentation methods to generate multi-scale cutting results, extracting multiple segmentation results from low to high thresholds, and establishing a multi-scale hierarchical region merging tree; and Quantify the segmentation quality of the segmented areas of each scale, and use the optimization algorithm to synthesize the global image hierarchy; determine the set of suitable segmented areas according to the result of the synthesis level, and generate the optimal segmentation result of the image according to the set. The specific steps are:

[0056] Step 1: Select an effective segmentation result from the multi-scale segmentation results.

[0057] Step 1.1: Use the multi-scale image segmentation method to obtain the segmentation result of each image in the database, that is, the ultrametric contour map (Ultrametric Contour Map, UCM), expan...

Embodiment 2

[0086] In order to verify the effectiveness of this application for multi-scale segmentation method scale perception, the method described in this application carried out scale perception experiments on BSDS500 and PACAL Context datasets, and used SegmentationCovering (SC), Probabilistic Rand Index (PRI), Variation of The three indicators of Information (VI) are used as image segmentation quality evaluation indicators to verify the effect of scale perception. The detailed information of the dataset is shown in Table 1.

[0087] Table 1 Description of related image segmentation database

[0088]

Embodiment 3

[0090] The method of the present invention uses five open-source multi-scale image segmentation methods that are currently popular and perform well to perform scale-aware experiments. Algorithms include: PMI, UCM, SCG, MCG and COB. In the actual test, five commonly used algorithms select the parameter values ​​with the best performance on the entire data set, and use the threshold with the best effect to extract the segmentation results. The method of the invention is used to sense the scale of five commonly used methods, and the optimal segmentation scale is synthesized from multiple scales. The comparison of the results of the method of the present invention with the original method is shown in Table 2. In Table 2, the higher the value of the indicators SC and PRI, the higher the quality of the segmentation, on the contrary, the smaller the value of VI, the higher the quality of the segmentation.

[0091] Table 2 shows the segmentation evaluation results of five multi-scal...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention relates to the field of multi-scale image segmentation, in particular to a multi-scale image segmentation method based on hierarchical region merging. The method comprises the steps of generating a multi-scale segmentation result by using a multi-scale image segmentation method, extracting a plurality of segmentation results from low to high according to a threshold value, and establishing a multi-scale hierarchical region merging tree; carrying out segmentation quality quantification on the segmentation area of each scale, and carrying out global image hierarchical synthesis byutilizing an optimization algorithm; and determining a set of suitable segmentation regions according to a result of the synthesis hierarchy, and generating an optimal segmentation result of the imageaccording to the set. According to the method, the segmentation precision loss caused by manually setting a threshold in a multi-scale image segmentation method is overcome, and the optimal segmentation scale selection of an individual target can be realized to a certain extent.

Description

technical field [0001] The invention relates to the field of multi-scale image segmentation, in particular to a multi-scale image segmentation method based on hierarchical region merging. Background technique [0002] Image segmentation is to divide an image into several non-overlapping sub-regions, so that the internal features of the regions are similar and the inter-regional features have obvious differences. It is a basic task in computer vision. Image segmentation divides the pixels of an image into distinct blocks, each representing a discriminative transaction in the image. Represent the image as a collection of physically meaningful connected regions, and mark and locate the target and background in the image according to the prior knowledge of the target and background, and then separate the target from the background or other pseudo-targets. [0003] Multi-scale image segmentation is one of the image segmentation methods, which can obtain tree-shaped segmentation ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06T7/11G06T7/136G06T7/44G06T7/90
CPCG06T7/11G06T7/136G06T7/44G06T7/90Y02D10/00
Inventor 彭博马小军李天瑞
Owner SOUTHWEST JIAOTONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products