Medical Image Fusion Based on Fast Finite Shearlet Transform and Sparse Representation

Abstract

Clinical diagnosis has high requirements for the visual effect of medical images. To obtain rich detail features and clear edges for fusion medical images, an image fusion algorithm FFST-SR-PCNN based on fast finite shearlet transform (FFST) and sparse representation is proposed, aiming at the problem of poor clarity of edge details that is conducive to maintaining the details of source image in current algorithms. Firstly, the source image is decomposed into low-frequency coefficients and high-frequency coefficients by FFST. Secondly, the K-SVD method is used to train the low-frequency coefficients to obtain the overcomplete dictionary D, and then the OMP algorithm sparsely encodes the low-frequency coefficients to complete the fusion of the low-frequency coefficients. Then, a high-frequency coefficient is applied to excite a pulse-coupled neural network, and the fusion coefficient of the high-frequency coefficient is selected according to the number of ignitions. Finally, the fused low-frequency coefficient and high-frequency coefficient are reconstructed into the fused medical image by FFST inverse transform. The experimental results show that the image fusion result of the proposed algorithm is about 35% higher than the comparison algorithms for the edge information transfer factor QAB/F index and has achieved good results in both subjective visual effects and objective evaluation indicators.

Evaluation on Effect of Acupoint Application to Treat Idiopathic Edema of Perimenopausal Women Using the Segmentation Dictionary Learning Algorithm

“…The S-DL algorithm extracted the main features from a large number of image blocks, so as to train a dictionary suitable for such image feature selection. The sparse representation means that only a few function bases in the complete dictionary, that is, a sparse atomic dictionary, can be used to represent the target object [ 16 , 17 ]. It was supposed that the sample data set M =[ m 1 , m 2 , m 3 …, m y ] ∈ P x × y ( x and y were the spatial dimensions of the sample), K ={ k 1 , k 2 , k 3 …, k d } ∈ P x × d , and x < d ( d was the number of function bases of the dictionary) and the constraint condition k d R k d ≤ 1 were satisfied; then the linearity of the dictionary can be expressed as …”

“…In equations ( 2 ) and ( 3 ), F u represented the Fourier transform and under-harvesting, n represented the k-space data obtained from under-harvesting, and ν represented the weight constant. The dictionary update uses the K-SVD [ 16 ] algorithm, and the sparse coefficient update uses the OMP [ 17 ] algorithm.…”

“…e S-DL algorithm extracted the main features from a large number of image blocks, so as to train a dictionary suitable for such image feature selection. e sparse representation means that only a few function bases in the complete dictionary, that is, a sparse atomic dictionary, can be used to represent the target object [16,17] (1)…”

Multimodal Medical Image Fusion Based on Multiple Latent Low-Rank Representation

“…However, since the SR-based algorithm's dictionary matrix cannot fully include source image data, it fails to extract the source image's detailed texture information. Some scholars applied MST [24][25][26] or filter [27][28][29] to decompose the source images. And SR can be used to fuse the low-frequency subbands.…”

Infrared and visible image fusion using multi‐scale NSCT and rolling‐guidance filter

“…Fast computational speed can be achieved by wavelet transform [11] and it gathers limited directional information while capturing edges and contour of an image along with the horizontal, vertical and diagonal directions, but it suffers from pseudo‐Gibbs phenomenon around singularities. In order to overcome these shortcomings, several multi‐scale transforms like shearlet [12, 13], curvelet [14], ripplet [3], contourlet [15] are introduced. All these aforementioned transforms suffer due to the lack of shift‐invariance.…”