Aiming at the problem that current deep learning based undersampled Magnetic Resonance (MR) image reconstruction methods mainly focus on the single slice reconstruction and ignore the data redundancy between adjacent slices, a Hybrid Cascaded Convolutional Neural Network (HC-CNN) was proposed for undersampled multi-slice brain MR image reconstruction. First, the traditional reconstruction method was extended to a deep learning based reconstruction model, and the traditional iterative reconstruction framework was replaced by a cascaded convolutional neural network. Then, in each iterative reconstruction, a 3D convolution module and a 2D convolution module were used to learn the data redundancy between adjacent slices and inside a single slice, respectively. Finally, Data Consistency (DC) module was used in each iteration to maintain the data fidelity of the reconstructed image in k-space. The simulation results on a single-coil brain MR image dataset show that compared with the reconstruction methods based on single slice reconstruction, the proposed method has the Peak Signal-to-Noise Ratio (PSNR) value at 4×acceleration factor increased by 1.75 dB averagely and the PSNR value at 6×acceleration factor increased by 2.57 dB averagely. At the same time, the image reconstruction time for a single slice by the proposed method is 15.4 ms. Experimental results show that the proposed method can not only effectively utilize the data redundancy between slices and reconstruct higher-quality images, but also has a higher real-time performance.

Considering the low accuracy of the existing image segmentation method based on affinity propagation clustering, a FCAP algorithm which combined fuzzy connectedness and affinity propagation clustering was proposed. A Whole Fuzzy Connectedness (WFC) algorithm was also proposed with concerning the shortcoming of traditional fuzzy connectedness algorithms that can not get fuzzy connectedness of every pair of pixels. In FCAP, the image was segmented by using super pixel technique. These super pixels could be considered as data points and their fuzzy connectedness could be computed by WFC. Affinities between super pixels could be calculated based on their fuzzy connectedness and spatial distances. Finally, affinity propagation clustering algorithm was used to complete the segmentation. The experimental results show that FCAP is much better than the methods which use affinity propagation clustering directly after getting super pixels, and can achieve competitive performance when comparing with other unsupervised segmentation methods.