No abstract

Compressed sensing MRI (CS-MRI) aims to significantly reduce the measurements required for image reconstruction in order to accelerate the overall imaging speed. The sparsity of the MR images in transformation bases is one of the fundamental criteria for CS-MRI performance. Sparser representations can require fewer samples necessary for a successful reconstruction or achieve better reconstruction quality with a given number of samples. Generally, there are two kinds of 'sparsifying' transforms: predefined transforms and data-adaptive transforms. The predefined transforms, such as the discrete cosine transform, discrete wavelet transform and identity transform have usually been used to provide sufficiently sparse representations for limited types of MR images, in view of their isolation to the object images. In this paper, we present singular value decomposition (SVD) as the data-adaptive 'sparsity' basis, which can sparsify a broader range of MR images and perform effective image reconstruction. The performance of this method was evaluated for MR images with varying content (for example, brain images, angiograms, etc), in terms of image quality, reconstruction time, sparsity and data fidelity. Comparison with other commonly used sparsifying transforms shows that the proposed method can significantly accelerate the reconstruction process and still achieve better image quality, providing a simple and effective alternative solution in the CS-MRI framework.

A practical convolutional neural network (CNN) model is proposed to discriminate the Raman spectra of human and animal blood. The proposed network, which discards the pooling layers to avoid loss of data, consists of preprocessing and fully connected classifier layers. Two preprocessing layers, namely, denoising and baseline correction layer, are designed to allow only one kernel for each layer to explicitly suppress the noise and subtract varying background of the spectra. The network combines the preprocessing and discrimination to form a whole processing unit and learns parameters adaptively by training from 217 of 326 Raman spectra of human, dog, and rabbit blood samples. The trained network is evaluated by remaining 109 samples and shows better classification accuracy, as compared with the PLSDA and SVM.

No abstract

A new method based on the weighted fusion of multiple models is presented for wavelength selection in multivariate calibration of spectral data. It fuses the regression coefficients of multiple models with weights based on minimum mean square error to improve the accuracy and stability of the wavelength selection. To validate the performance of the proposed method, it was applied to the partial least squares (PLS) modeling of three near-infrared spectral datasets and compared with full-spectrum PLS, genetic algorithm-based PLS, and uninformative variable elimination-based PLS methods. Results show that the proposed method can effectively select the informative wavelength and enhance the prediction ability of the PLS model. On account of its simpler algorithm and higher efficiency, it can be widely used in practical applications.