Image semantic segmentation optimization method and device, storage medium and terminal

A technology of semantic segmentation and optimization method, which is applied in the field of computer vision and can solve the problems of poor segmentation effect and difficulty in obtaining image pixel dependencies.

Pending Publication Date: 2019-06-11
江苏通祐机器人科技有限公司
View PDF5 Cites 15 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Since the above convolution process is space-invariant, it obtains the relationship between image regions and regions, and it ...

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
  • Image semantic segmentation optimization method and device, storage medium and terminal
  • Image semantic segmentation optimization method and device, storage medium and terminal
  • Image semantic segmentation optimization method and device, storage medium and terminal

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0029] figure 1 It is a flow chart of an image semantic segmentation optimization method provided in Embodiment 1 of the present application. The method can be executed by an image semantic segmentation optimization device, wherein the device can be implemented by software and / or hardware, and can generally be integrated in in the smart terminal. Such as figure 1 As shown, the method includes:

[0030] Step 110, acquiring superpixels in the image to be segmented.

[0031] For example, a superpixel (that is, a superpixel) is a continuous, non-overlapping area composed of a series of adjacent pixels in the image to be segmented and having similar characteristics such as brightness, color, and texture. Wherein, the image to be segmented is the original image to be subjected to image semantic segmentation. The color mode of the image to be segmented may be RGB, or other color modes. It should be noted that the color is usually described by three relatively independent attribu...

Embodiment 2

[0074] image 3 It is a flow chart of an optimization method for image semantic segmentation provided in Embodiment 2 of the present application. This embodiment further refines the above-mentioned steps related to image semantic segmentation. Such as image 3 As shown, the method includes:

[0075] Step 301. Acquire an image to be segmented for image semantic segmentation.

[0076] Exemplarily, the image to be segmented may be an RGB image, or an image in other color modes. Wherein, the width and height of the RGB image may be w*h.

[0077] Step 302. Input the image to be segmented into the convolutional neural network model.

[0078] Among them, the convolutional neural network model can be a full convolutional neural network model. The full convolutional neural network model has no restrictions on the input image, receives an input of any size, and calculates an output image of a semantic segmentation result. The split images are of the same size. The fully convolutio...

Embodiment 3

[0109] Figure 8 A structural block diagram of an image semantic segmentation optimization device provided in Embodiment 3 of the present application. The device can be implemented by software and / or hardware, and is generally integrated in a smart terminal, and image semantics can be optimized by executing an image semantic segmentation optimization method Split results. Such as Figure 8 As shown, the device includes:

[0110] A superpixel acquisition module 810, configured to acquire superpixels in the image to be segmented;

[0111] The distribution information determination module 820 is configured to obtain a probability map of the image to be segmented, and determine probability distribution information of the label category of the superpixel according to the probability map, wherein the probability map is used to represent The probability of the label category of each pixel in the image to be segmented;

[0112] A superpixel adjustment module 830, configured to adj...

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 embodiment of the invention discloses an image semantic segmentation optimization method and device, a storage medium and a terminal. The method comprises the steps of obtaining superpixels in a to-be-segmented image; obtaining a probability mapping graph of the to-be-segmented image, and determining probability distribution information of a label class of the superpixels according to the probability mapping graph; determining a target superpixel meeting a set condition according to the probability distribution information, and adjusting a label class of a pixel point in the target superpixel; constructing a conditional random field model based on the adjusted target superpixel and the remaining superpixels in the to-be-segmented image, and determining an image semantic segmentation result of the to-be-segmented image according to the conditional random field model, thereby realizing optimization of the segmentation result of the to-be-segmented image according to the conditional random field, and improving the segmentation effect of the image semantic segmentation result at the boundary.

Description

technical field [0001] The embodiments of the present application relate to the field of computer vision, and in particular to an optimization method, device, storage medium and terminal for image semantic segmentation. Background technique [0002] As the cornerstone technology of image understanding, image semantic segmentation plays a pivotal role in many aspects, such as autonomous driving, drone applications, wearable devices and so on. [0003] At the pixel level, image segmentation is to assign an object category label to each pixel in the image. Pixels of the same category label are aggregated into a region, representing the same local representation, usually with clear semantic information, that is, representing objects or parts of objects. Research methods for image semantic segmentation at home and abroad are mainly divided into: feature-based methods and deep learning-based methods. There have been decades of research around feature-based methods. This method i...

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): G06T7/90G06N3/04G06K9/62
Inventor 王琰张亮朱光明刘挺
Owner 江苏通祐机器人科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap