A Discrete Asynchronous Event Data Augmentation Method

A technology of asynchronous events and event data, applied in the field of data processing, can solve the problem of low generalization ability of deep learning models, and achieve the effect of improving generalization ability, increasing diversity, and improving diversity

Active Publication Date: 2022-05-27
CHONGQING UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

It is used to solve the key problem of how to enhance the event data to overcome the low generalization ability of the deep learning model caused by the noise and occlusion problems in the event data, and to expand the application of the event data-oriented deep learning model

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 Discrete Asynchronous Event Data Augmentation Method
  • A Discrete Asynchronous Event Data Augmentation Method
  • A Discrete Asynchronous Event Data Augmentation Method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the preferred embodiments are only for illustrating the present invention, rather than for limiting the protection scope of the present invention.

[0030] The present invention proposes three methods for enhancing event data, including random discarding, time discarding and regional discarding of a certain proportion of data. The random drop method is used to overcome the noise problem of event data, and the other two strategies are used to simulate different occlusion situations. figure 1 The idea of ​​different event data augmentation strategies is illustrated. exist figure 1 where t represents the time dimension, x represents the pixel coordinates (only one dimension is shown here for clarity), the black dots represent the original events, the dots inside the rectangular dashed box represent the events to be...

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 discrete asynchronous event data enhancement method, which includes the following steps: Step S1: input the asynchronous event stream and the corresponding image resolution as event data; Step S2: initialize variables; The data is enhanced; step S4: output the enhanced event data. This method effectively increases the amount of event data and the diversity of event data by adopting random discarding, discarding by time and discarding by region operations, and to a certain extent solves the problems caused by noise and occlusion in event data. model overfitting problem. The method proposed by the invention is easy to implement and has low calculation cost, can significantly improve the generalization performance of various models including deep learning models, and can be widely used in various event-based learning tasks.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a discrete asynchronous event data enhancement method. Background technique [0002] Event-based data often have some characteristics that have an important impact on the generalization ability of deep learning models. For example, the output of an event camera can vary significantly over time under the same lighting conditions and scene. This is mainly due to random noise in the event camera data. Randomly removing a portion of event data can improve the diversity of event data, thereby improving the performance of downstream applications. [0003] Furthermore, in many tasks such as object recognition and tracking, event data may suffer from occlusion issues. However, the generalization ability of machine learning models largely depends on the diversity of training data, including data in various occluded situations. However, the available training data usually allow...

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): G06N3/08G06N3/04
CPCG06N3/08G06N3/045
Inventor 古富强余芳文胡旭科
Owner CHONGQING UNIV
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