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

Raw domain video denoising supervision data set construction method

A construction method and technology of video data, applied in image data processing, neural architecture, instruments, etc., can solve problems such as the lack of raw domain video denoising data sets

Inactive Publication Date: 2020-09-29
TIANJIN UNIV
View PDF4 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, methods based on convolutional neural networks often require a large amount of data for training. At present, there are many public real image denoising datasets, but there is no Raw domain video denoising dataset in dynamic scenes.

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
  • Raw domain video denoising supervision data set construction method
  • Raw domain video denoising supervision data set construction method
  • Raw domain video denoising supervision data set construction method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] Specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0024] A kind of Raw domain video denoising supervisory dataset construction method of the present invention, specifically comprises the following steps:

[0025] Step 1. Model the Raw domain noise as a Poisson-Gaussian mixed noise model, and the noise model expression is:

[0026]

[0027] where x p Indicates the observed noise image, y p represents an ideal noise-free image. Indicates the variance of the Gaussian noise component, Represents the variance of the Poisson noise component. and Indicates the parameters of the noise model, which change with the change of ISO, represents a Poisson distribution, Indicates a Gaussian distribution, ~ indicates that it belongs to the distribution of ...;

[0028] Step 2. Shoot Flat-field frames and Bias frames, and correct the noise model with the parameters under the specified ISO;

...

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 Raw domain video denoising supervision data set construction method. The method comprises the steps that 1, modeling Raw domain noise into Poisson-Gaussian mixed noise; 2, shooting a Flat-field frame and a Bias frame, and correcting parameters of the noise model under a specified ISO; 3, performing ISP simulation and inverse ISP simulation, visualizing a raw image, generating a corresponding sRGB domain video image, and converting the sRGB domain video image into a raw domain video image; and 4, preparing data sets, including a synthetic noise simulation data set anda real Raw domain video data set. Compared with the prior art, the Raw domain video denoising supervision data set construction method can be used for simulating Raw domain noise, can be used for constructing a large amount of clean-noise paired Raw domain video data, and supports Raw domain video denoising work.

Description

technical field [0001] The invention belongs to the technical field of computer image processing, and in particular relates to a related processing technology of Raw domain video denoising. Background technique [0002] In recent years, with the rapid development of smart phones, surveillance cameras, and autonomous driving, people's demand for high-quality images and videos has become increasingly prominent. However, it is difficult to ensure high-quality imaging in harsh environments. For example, most imaging devices tend to set higher ISOs under low-light conditions, resulting in a lot of noise in the generated video, which affects subsequent video analysis and viewers. visual experience. On the Raw domain video directly recorded by the sensor, the noise follows a simple distribution pattern close to the Poisson Gaussian distribution, but on the sRGB domain video processed by the imaging device ISP (Image Signal Processor), the noise distribution becomes very complicate...

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 Applications(China)
IPC IPC(8): G06T5/00G06T5/50G06N3/04
CPCG06T5/50G06T2207/10016G06N3/045G06T5/70
Inventor 岳焕景曹聪杨敬钰
Owner TIANJIN 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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com