A Saliency Detection Method Based on Multiscale Wavelet Transform of Discrete Cosine Coefficients

Abstract

The invention discloses an image visual saliency calculation method based on discrete cosine coefficient multi-scale wavelet transform, which belongs to the technical field of image processing. The method includes scaling the input image size; calculating the generalized red, green, and blue color channels and the intensity channel; calculating the weight coefficients of the red, green, blue, and intensity channels; calculating the magnitude of the discrete cosine transform of the red, green, blue, and intensity channels Matrices and symbolic matrices; Compute multiscale magnitude matrices for red, green, blue, and intensity channels using multiscale wavelet transform; Compute multiscale channel saliency maps for red, green, blue, and intensity channels; Synthesize multiscale channel saliency maps Spatial domain visual saliency map at multiple scales; select a better saliency map according to the saliency evaluation function to generate a fused saliency map; perform central bias optimization on the fused saliency map to generate the final saliency map. The invention can quickly and effectively calculate the saliency value of the image, and the obtained saliency image has complete saliency objects and less background interference.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a saliency detection method based on discrete cosine coefficient multi-scale wavelet transform. Background technique [0002] The human visual system has a bottom-up visual attention mechanism, which can quickly grasp and extract objects of interest, greatly reducing the occupation of brain nerve resources. The selective visual attention mechanism processes information in a serial manner of neurons, allowing only a small amount of perceptual information to enter the visual higher cortex, thereby highlighting salient objects and ignoring surrounding background areas. According to the attention mechanism, the visual attention mechanism is divided into scene-dependent or bottom-up visual attention and task-dependent or top-down visual attention. The visual saliency detection algorithm simulates this visual attention mechanism by computer, highlights the salient target, sup...

AI Technical Summary

Problems solved by technology

[0005] Recently, a class of saliency detection methods based on deep learning has emerged. These methods include "Deep visual attention prediction" proposed by Wang et al. in 2017 and "Efficientsaliency detection using convolutional neural networks with feature selection" proposed by Cao et al. in 2018. The saliency map obtained by the class method has high accuracy, but the training process is complicated and the calculation speed is slow

Images