An Image Blur Detection Method Based on Saliency Detection

Abstract

The invention discloses an image blur detection method based on saliency detection. It collects sample images, first performs fast Fourier transform and size reduction processing on the RGB three channels of the most salient area of ​​each sample image, and uses the size The reduced complex data constructs a new image, and the newly constructed reduced image is used as the input image of the convolutional neural network, and the clear-fuzzy image classification training is performed to obtain the blur detection model, and finally the most significant area of ​​the image to be detected is analyzed Fast Fourier transform and size reduction processing of the three RGB channels, and use the reduced complex data to construct a new image to be detected, and then use the blur detection model to perform clear-blur on the newly constructed image to be detected Image discrimination, thereby effectively reducing the amount of calculation and speeding up the detection speed. It is especially suitable for fast blur detection of large-size images and improves the accuracy of detection.

Description

technical field [0001] The invention relates to an image processing method, in particular to an image blur detection method based on saliency detection. Background technique [0002] Digital image processing has become a basic research object in many fields such as information science, biology, and medicine. With the advent of the information age, digital image processing has been widely used in computer vision, machine learning, artificial intelligence and other fields, and its importance has become increasingly prominent. Regrettably, in the process of digital image acquisition, the acquisition device will produce unavoidable slight jitter at the moment the shutter is opened. This jitter often makes us only get an image with blurred details, especially in the light In less-than-ideal conditions, longer shutter times can result in more blurred images. Such fuzzy images have brought great troubles to human vision, and at the same time lost a lot of detailed information, wh...

AI Technical Summary

Problems solved by technology

Regrettably, in the process of digital image acquisition, the acquisition device will produce unavoidable slight jitter at the moment the shutter is opened. This jitter often makes us only get an image with blurred details, especially in the light In less than ideal conditions, longer shutter times make images more blurred

Such fuzzy images have brought great troubles to human vision, and at the same time lost a lot of detailed information, which cannot be applied in daily life and scientific research activities

The existing image blur detection methods can be roughly divided into two categories: one class gives the estimation of the blur degree of the whole image, and the other class divides the image into several regions and gives estimates of the blur degree for each region respectively. However, most of the calculation methods are more complex, the amount of calculation is large, and the processing speed is slow.

Images