Sparse subspace clustering via group sparse coding

We propose in this paper a novel sparse subspace clustering method that regularizes sparse subspace representation by exploiting the structural sharing between tasks and data points via group sparse coding. We derive simple, provably convergent, and computationally efficient algorithms for solving the proposed group formulations. We demonstrate the advantage of the framework on three challenging benchmark datasets ranging from medical record data to image and text clustering and show that they consistently outperforms rival methods.

Sparse coding for improved signal-to-noise ratio in MRI

Magnetic Resonance images (MRI) do not only exhibit sparsity but their sparsity take a certain predictable shape which is common for all kinds of images. That region based localised sparsity can be used to de-noise MR images from random thermal noise. This paper present a simple framework to exploit sparsity of MR images for image de-noising. As, noise in MR images tends to change its shape based on contrast level and signal itself, the proposed method is independent of noise shape and type and it can be used in combination with other methods.

Object categorization via sparse representation of local features

Sparse representation has been introduced to address many recognition problems in computer vision. In this paper, we propose a new framework for object categorization based on sparse representation of local features. Unlike most of previous sparse coding based methods in object classification that only use sparse coding to extract high-level features, the proposed method incorporates sparse representation and classification into a unified framework. Therefore, it does not need a further classifier. Experimental results show that the proposed method achieved better or comparable accuracy than the well known bag-of-features representation with various classifiers.

Human identification from ECG signals via sparse representation of local segments

This work proposes a novel framework to extract compact and discriminative features from Electrocardiogram (ECG) signals for human identification based on sparse representation of local segments. Specifically, local segments extracted from an ECG signal are projected to a small number of basic elements in a dictionary, which is learned from training data. A final representation is extracted by performing a max pooling procedure over all the sparse coefficient vectors in the ECG signal. Unlike most of existing methods for human identification from ECG signals which require segmentation of individual heartbeats or extraction of fiducial points, the proposed method does not need to segment individual heartbeats or detect any fiducial points. The method achieves an 99.48% accuracy on a 100 subjects dataset constructed from a publicly available database, which demonstrates that both local and global structural information are well captured to characterize the ECG signals.

Sparse Representations of Image via Overcomplete Dictionary Learned by Adaptive Non-orthogonal Sparsifying Transform

How to learn an over complete dictionary for sparse representations of image is an important topic in machine learning, sparse coding, blind source separation, etc. The so-called K-singular value decomposition (K-SVD) method [3] is powerful for this purpose, however, it is too time-consuming to apply. Recently, an adaptive orthogonal sparsifying transform (AOST) method has been developed to learn the dictionary that is faster. However, the corresponding coefficient matrix may not be as sparse as that of K-SVD. For solving this problem, in this paper, a non-orthogonal iterative match method is proposed to learn the dictionary. By using the approach of sequentially extracting columns of the stacked image blocks, the non-orthogonal atoms of the dictionary are learned adaptively, and the resultant coefficient matrix is sparser. Experiment results show that the proposed method can yield effective dictionaries and the resulting image representation is sparser than AOST.

Extremely high conservation in the untranslated Region as well as the coding region of CNP mRNAs throughout elasmobranch species

C-type natriuretic peptide (CNP) is a crucial osmoregulatory hormone in elasmobranchs, participating in salt secretion and drinking. In contrast to teleosts and tetrapods in which the NP family is composed of a group of structurally related peptides, we have shown that CNP is the sole NP in sharks. In the present study, CNP cDNAs were cloned from four species of batoids, another group of elasmobranchs. The cloned batoid CNP precursors contained a plausible mature peptide of 22 amino acid residues that is identical to most shark CNP-22s, but five successive amino acids were consistently deleted in the prosegment compared with shark precursors, supporting the diphyletic classification of sharks and rays. In addition, molecular phylogenetic trees of CNP precursors were consistent with a diphyletic interpretation. Except for the deletion, the nucleotide and deduced amino acid sequences of the CNP cDNAs are extremely well-conserved among all elasmobranch species, even between sharks and rays. Surprisingly, high conservation is evident not only for the coding region, but also for the untranslated regions. It is most likely that the high conservation is due to the low nucleotide substitution rate in the elasmobranch genome, and high selection pressure. The 3′-untranslated region of the elasmobranch CNP cDNAs contained three to six repeats of the ATTTA motif that is associated with the regulation of mRNA stability and translation efficiency. Alternative polyadenylation sites were also found; the long 3′-untranslated region contains a core of ATTTA motifs while the short form has only one or no ATTTA motif, indicating that the post-transcriptional modification of mRNA is important for regulation of CNP synthesis. These characteristics in the 3′-untranslated region were conserved among all elasmobranch CNP cDNAs. Since CNP has been implicated as a fast-acting hormone to facilitate salt secretion from the rectal gland, the conserved 3′-untranslated region most likely contributes to rapid regulation of CNP synthesis in elasmobranchs in response to acute changes in internal and external environments.

Massive data delivery in unstructured peer-to-peer networks with network coding

With more and more multimedia applications on the Internet, such as IPTV, bandwidth becomes a vital bottleneck for the booming of large scale Internet based multimedia applications. Network coding is recently proposed to take advantage to use network bandwidth efficiently. In this paper, we focus on massive multimedia data, e.g. IPTV programs, transportation in peer-to-peer networks with network coding. By through study of networking coding, we pointed out that the prerequisites of bandwidth saving of network coding are: 1) one information source with a number of concurrent receivers, or 2) information pieces cached at intermediate nodes. We further proof that network coding can not gain bandwidth saving at immediate connections to a receiver end; As a result, we propose a novel model for IPTV data transportation in unstructured peer-to-peer networks with network coding. Our preliminary simulations show that the proposed architecture works very well.

Chemical coding of vascular nitrergic nerves in lower vertebrates

Nitric oxide is an important regulator of blood pressure in mammals. This study provided new information on the role of nitric-oxide releasing sympathetic nerves in vascular regulation of lower vertebrates. The research outcomes advance knowledge on the potential role of these unique nerves in the control of the lower vertebrate cardiovascular system.