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

Target detection method based on improved RetinaNet microscopic image

A target detection algorithm and microscopic image technology, applied in image analysis, image enhancement, image data processing, etc., can solve the problems of heavy workload, long detection time, and inability to realize multi-target detection requirements, etc. high precision effect

Active Publication Date: 2020-07-14
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the existing methods based on machine learning and deep learning have low detection accuracy, long detection time, and poor real-time performance. The detected samples need to be reviewed by doctors, and t

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
  • Target detection method based on improved RetinaNet microscopic image
  • Target detection method based on improved RetinaNet microscopic image
  • Target detection method based on improved RetinaNet microscopic image

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] Below in conjunction with accompanying drawing, the type component detection method of microscopic image in the present invention is described in detail:

[0035] Step 1: Use a microscopic imaging system to collect microscopic images of leucorrhea samples, and select the three clearest images in each field of view as the sample set;

[0036] Step 2: Manually mark the image collected in step 1, and mark the position and type of the shaped components;

[0037] Step 3: Build a RetinaNet convolutional neural network model, such as figure 2 shown;

[0038] Step 3-1: Build a RetinaNet network structure model, use ResNet-50 as the feature extraction layer of the network, and generate a feature pyramid network;

[0039] Step 3-2: Since the stylish components in the leucorrhea sample contain 6 categories, the output of the network is adjusted accordingly.

[0040] Step 4: Generate anchor point information for the feature maps of each layer of the model, such as image 3 sho...

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 discloses a target detection method based on an improved RetinaNet microscopic image. The method is mainly applied to detection items related to cell microscopic examination in the clinical laboratory of a hospital. According to the algorithm, an anchor point extraction mode, classification and a regression sub-network in RetinaNet are improved, and high-efficiency and high-accuracydetection of microscopic image morphological components is realized. The algorithm involved in the invention provides a theoretical basis for realizing full automation and intelligence of detection. The detection method is high in speed and high in detection precision, and can completely meet the requirements of clinical detection.

Description

technical field [0001] The invention is an automatic positioning and identification algorithm applied to cells and other shaped components in microscopic images, and its detection model is based on an improved RetinaNet model. Background technique [0002] The positioning and identification technology of visible components in microscopic images is widely used in hospital laboratory departments, such as stool routine, leucorrhea routine, urine routine and other inspection items. At present, the detection methods of shaped components in microscopic images mainly rely on manual interpretation. In recent years, with the development of machine learning and deep learning, recognition technology has gradually become more automated and intelligent. However, the existing methods based on machine learning and deep learning have low detection accuracy, long detection time, and poor real-time performance. The detected samples need to be reviewed by doctors, and the workload is heavy, w...

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): G06K9/00G06K9/62G06T7/00G06N3/04G06N3/08
CPCG06T7/0012G06N3/08G06T2207/10056G06T2207/20081G06T2207/30024G06V20/693G06V20/695G06V20/698G06N3/045G06F18/23213
Inventor 杜晓辉刘霖张静王祥舟郝如茜倪光明刘娟秀刘永
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA