Unlock instant, AI-driven research and patent intelligence for your innovation.

Object Semantic Segmentation Method and Its Applied Street Object Anomaly Detection Method

A semantic segmentation and target technology, applied in the field of target detection, can solve problems such as difficult segmentation of small targets, difficulty in labeling data, and inaccurate segmentation of large targets

Active Publication Date: 2022-04-22
CITY CLOUD TECH HANGZHOU CO LTD
View PDF12 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The current semantic segmentation obviously has problems such as difficulty in labeling data, difficulty in segmenting small objects, and inaccurate segmentation of large objects.

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
  • Object Semantic Segmentation Method and Its Applied Street Object Anomaly Detection Method
  • Object Semantic Segmentation Method and Its Applied Street Object Anomaly Detection Method
  • Object Semantic Segmentation Method and Its Applied Street Object Anomaly Detection Method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] This embodiment provides a target semantic segmentation method for performing target semantic segmentation on an image to be detected. Such as figure 1 As shown, the method can be summarized into 3 steps:

[0052] Step 1, use the region growing algorithm to segment the target in the image to be detected, and obtain the segmentation contour of each target in the image to be detected; Step 2, input the image to be detected into the segmentation network to obtain the target category probability of each pixel, according to each The target category probability of one pixel is used to obtain the predicted contour of each target; step 3, according to the segmentation contour and predicted contour corresponding to the same target, the semantic segmentation result of the target is obtained.

[0053] In step 1, first obtain the image to be detected, input the image to be detected into the target detection network for target detection, obtain at least one detection frame, then de...

Embodiment 2

[0086] Based on the same idea, this embodiment also provides a target semantic segmentation device for implementing the target semantic segmentation method described in Embodiment 1. The device includes the following modules:

[0087] a first acquiring module, configured to acquire an image to be detected including at least one target;

[0088] A target detection module, configured to input the image to be detected into a target detection network to obtain at least one detection frame of the target;

[0089] A region growing module, configured to determine at least one growth region according to each of the detection frames, select an initial seed point in each of the growth regions to perform region growth, and obtain a corresponding The segmentation contour of the target;

[0090] The first segmentation module is used to input the image to be detected into the target segmentation network to obtain the target category probability of each pixel, and determine each pixel whose...

Embodiment 3

[0097] This embodiment also provides an electronic device, refer to Figure 4 , including a memory 404 and a processor 402, wherein a computer program is stored in the memory 404, and the processor 402 is configured to run the computer program to perform the steps of any one of the object semantic segmentation method or street object anomaly detection method in the above-mentioned embodiments .

[0098] Specifically, the processor 402 may include a central processing unit (CPU), or an Application Specific Integrated Circuit (ASIC for short), or may be configured to implement one or more integrated circuits in the embodiments of the present application.

[0099]Wherein, the memory 404 may include a mass memory 404 for data or instructions. By way of example and not limitation, the memory 404 may include a hard disk drive (Hard Disk Drive, referred to as HDD), a floppy disk drive, a solid state drive (Solid State Drive, referred to as SSD), flash memory, optical disk, magneto-o...

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 present application proposes a target semantic segmentation method, including: acquiring an image to be detected including at least one target; inputting the image to be detected into a target detection network to obtain a detection frame of at least one target; setting a growth area and an initial seed point according to the detection frame, And according to the region growing results of each initial seed point, the segmentation contour of the corresponding target is obtained; the image to be detected is input into the target segmentation network to predict the contour of each target; Semantic segmentation results of the target. This method solves the problem of difficult labeling in semantic segmentation by using target detection to generate target detection frames, and determines the growth area and initial seed points based on the detection frame. Using the region growing algorithm to segment the target can obtain a more accurate segmentation outline of the target, and using The object segmentation network obtains the predicted contour of the object, and the semantic segmentation results are obtained from the segmented contour and the segmented contour of the same object.

Description

technical field [0001] The present application relates to the technical field of target detection, in particular to a target semantic segmentation method and a street target anomaly detection method using the same. Background technique [0002] Semantic segmentation is to classify each pixel in an image, which is currently widely used in various fields. The usual semantic segmentation architecture can be broadly thought of as an encoder network and a decoder network: the encoder is usually a pre-trained classification network such as vgg / resnet, and the task of the decoder is to integrate the discriminative features learned by the encoder into (low-resolution) semantic projection onto pixel space (high-resolution), resulting in per-pixel classifications. [0003] The current semantic segmentation obviously has problems such as difficulty in labeling data, difficulty in segmenting small objects, and inaccurate segmentation of large objects. Contents of the invention [00...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/11G06T7/00G06K9/62G06T7/90G06V10/764G06V10/774G06V10/80G06V10/74
CPCG06T7/11G06T7/0002G06T7/90G06F18/22G06F18/214G06F18/2415G06F18/253
Inventor 徐志坚章东平陈斌雷羽文董墨江
Owner CITY CLOUD TECH HANGZHOU CO LTD