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

A ResNet-based small target recognition method based on eagle brain feature integration

A feature integration, small target technology, applied in the field of computer vision, can solve the problems of reduced image quality and unfavorable target recognition, and achieve the effect of reducing parameters and improving the classification effect.

Active Publication Date: 2019-02-15
BEIHANG UNIV
View PDF5 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The reduction of image quality is not conducive to target recognition

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
  • A ResNet-based small target recognition method based on eagle brain feature integration
  • A ResNet-based small target recognition method based on eagle brain feature integration
  • A ResNet-based small target recognition method based on eagle brain feature integration

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0026]The technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0027] The effectiveness of the method designed in the present invention is verified below through specific examples of small target recognition. The deep learning framework used in this example is PyTorch, which is a deep learning tensor library optimized by GPU and CPU. In this framework, an automatic derivation mechanism is provided, which is very flexible and fast, and is convenient for deep learning development. . The main hardware configuration of the server used is as follows: the central processor model is Intel Core six-core i7-6850K, the 4 graphics cards are all GTX1080Ti, each graphics card has 11GB of memory, and the server has 16GB of memory. In this example, 2 GPUs are used for training. During training, each batch includes 8 training images, and the entire training library is traversed 20 times. The overall pro...

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 resNet-based small target recognition method for imitating eagle brain feature integration: step 1, establishing a small target recognition map library; 2, initializing thatsetting; 3, normalize images: calculate that mean value and variance of all images in the small target recognition map library, normalizing all images by use the mean value and variance, and normalizing the image size; Step 4: Calculating ResNet-34 convolution lay output; 5, integrating that characteristics of the imitate eagle brain; 6, classifying that small targets; 7, training thesmall targetclassification network; 8, testing that small target classification network. The method of the invention can greatly reduce the parameters of the whole connection layer by using the global pooling operation, and the method of using the feature parallel fusion can simultaneously utilize the features of different levels to carry out the small target recognition, thereby achieving better classification effect.

Description

technical field [0001] The invention is a ResNet-based imitation eagle brain feature integration small target recognition method, which belongs to the technical field of computer vision. Background technique [0002] Among all animals, eagle eyes are among the best in observing animals, and are famous for their wide field of vision and keen eyesight. From the appearance point of view, the eagle's eyes are relatively round, characterized by flat crystals and far away from the retina that receives light. Therefore, the eagle's visual system has the characteristics of long focal length, and has the function of long and short focal length conversion, which further enhances its sensitivity. But usually the image is darker because the eagle's eyes don't receive enough light. To overcome this potential disadvantage, eagles improved by having large pupils that flood the eyes with light. In addition, as an animal that relies on keen vision to capture prey information, the eagle can...

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): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/217G06F18/24G06F18/214
Inventor 段海滨王晓华邓亦敏李晗辛龙郭彦杰孙永斌徐小斌张锡联
Owner BEIHANG 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