Search

Music Source Separation (MSS) aims to separate the instrument sources from themixture. Since MUSDB18 is the standard dataset for MSS, most MSS systems aredeveloped on MUSDB18 and they can only focused on four stems, namely vocals,drums, bass, other. It is still impossible to separate more fine-grained level of sources,e.g., lead vocal, backing vocal and guitar. The limitation of MUSDB18 dataset andthe lack of large multi-track datasets are the possible obstacles. In this thesis, we first focus on building a new multi-track dataset called MS21,which can be viewed as an expanded and MSS oriented version of the Mixing Secretsdataset. With a total mixture duration of 34.2 hours, MS21 features multi-trackcontent in an unprecedented size and industrial standard. Statistics analysis showsthat MS21 contains more averaged number of multi-track files (20±13), a higherpercentage of vocal stem (84%) among multi-track dataset and a more diverse genredistribution among edleyDB, MUSDB18 and Slakh2100. MS21 provides MedleyDBlike metadata files which capture the specially designed three-level hierarchy annotation for all 500 songs, in order to bridge between different source separationresearch scenarios. Hierarchical Music Source Separation (HMSS) is a new research scenario, whichaims to separate the sources that share hierarchical relation, for example accompaniment- guitar - electric guitar. Multi-level HMSS system showed better performancewhen separating sources at more fine-grained level, compared to the singlelevelMSS system. In this thesis, to further explore the research possibility in HMSS, we proposed a theory called Mask Relation and a new mask-based method calledSequential Mask. Our HMSS experiments showed that the sequential mask-basedapproach provides worse results for higher-level sources, but better results for lowerlevel sources (+2dB SI-SDR) compared to a single-level MSS system, which we explain as the leaf-oriented feature of the sequential mask.

Introduction: Bone metabolism has an essential role in the bone disease, but its specific mechanism remains unclear. Y-Box Binding Protein 1 (YBX1) is a gene with broad nucleic acid binding properties, which encodes a highly conserved cold shock domain protein. Previous studies have shown that YBX1 is closely related to cell differentiation. However, the function of YBX1 in osteoblast differentiation of bone marrow mesenchymal stem cells (MSCs) was unclear. Method: To explore the effect and specific mechanism of YBX1 in osteogenic differentiation of MSCs, we used PCR, Western blot, Alizarin red Staining, alkaline phosphatase (ALP) assays, and siRNA knockdown in our research. We found that YBX1 gradually increased during the process of osteogenic differentiation of MSCs. YBX1 siRNA could negatively regulate the MSCs osteogenic differentiation. Mechanistic studies revealed that YBX1 knockdown could inhibit PI3K/AKT pathway. Furthermore, the specific agonist (SC79) of PI3K/AKT pathway could restore the impaired MSCs osteogenic differentiation which was mediated by YBX1 knockdown. Taken together, we concluded that YBX1 could positively regulate the osteogenic differentiation of MSCs by activating the PI3K/AKT pathway. Results and Discussion: These results helped us further understand the mechanism of osteogenesis and revealed that YBX1 might be a selectable target in the bone repair field. Conclusion: Our study provides a new target and theoretical basis for the treatment of bone diseases.

The multivariate time series (MTS) classification is an important classification problem in which data has the temporal attribute. Because relationships between many variables of the MTS are complex and time-varying, existing methods perform not well in MTS classification with many attribute variables. Thus, in this paper, we propose a novel model-based classification method, called Kullback-Leibler Divergence-based Gaussian Model Classification (KLD-GMC), which converts the original MTS data into two important parameters of the multivariate Gaussian model: the mean vector and the inverse covariance matrix. The inverse covariance is the most important parameter, which can obtain the information between the variables. So that the more variables, the more information could be obtained by the inverse covariance, KLD-GMC can deal with the relationship between variables well in the MTS. Then the sparse inverse covariance of each subsequence is solved by Graphical Lasso. Furthermore, the Kullback-Leibler divergence is used as the similarity measurement to implement the classification of unlabeled subsequences, because it can effectively measure the similarity between different distributions. Experimental results on classical MTS datasets demonstrate that our method can improve the performance of multivariate time series classification and outperform the state-of-the-art methods.

Background: To evaluate the feasibility of autologous cultivated oral mucosal epithelial transplantation (COMET) for the treatment of limbal stem cell deficiency (LSCD).Methods: Seven eyes from seven different patients with monocular LSCD were included in this study. Autologous oral mucosal epithelial cells were fabricated on ex vivo using amniotic membranes as a substrate. Clinical efficacy was evaluated by the coefficient of best-corrected visual acuity (BCVA). Clinical formation of the conjunctiva and symblepharon was evaluated and graded on a scale from 0 to 3. Clinical safety was evaluated by the presence of persistent epithelial defects, infection, and ocular hypertension.Results: Autologous COMET was successfully performed in all seven patients. The mean follow-up period was 10.7 months, during which time the postoperative formation of the conjunctiva and symblepharon was inhibited. BCVA was improved more than two lines in six eyes (86%) during the follow-up period. Complete reepithelialization of the corneal surfaces occurred in all treated eyes. No persistent epithelial defects, corneal infection, or postoperative ocular hypertension were observed. Conclusions: Autologous COMET offers a viable and safe alternative for the reconstruction of a stable ocular surface and improves vision in patients with LSCD.

Effects of (Lu, Ni) co-doping on spin entropy of Ca 3 Co 4 O 9+ δ have been investigated. Results of this study show that (Lu, Ni) co-doping can be more effective than single Lu doping for optimizing the spin entropy of Ca 3 Co 4 O 9+ δ . X-ray photo-emission spectroscopy confirms a reduction in Co 4+ concentration, which leads to enhanced spin entropy from Co sites. The magnetic ions of Ni with unpaired 3d electrons can produce extra spin entropy through a new hopping model, thereby contributing to an increase in total spin entropy. Incorporating magnetic ions is an effective way to optimize the spin entropy of thermoelectric materials. This study opens a new way to broaden and design prospective thermoelectric materials.

Rationale: Uveitis-glaucoma-hyphema (UGH) syndrome could be identified by conventional ultrasound biomicroscopy (UBM) and B-ultrasonography, but failed in some untypical cases. We introduced a novel application of B-ultrasonography in diagnosis of UGH syndrome in a rare case. Patient concerns: A 60-year-old woman was referred for distending pain with blurred vision for more than 1 month in the right eye after cataract surgery. Diagnoses: B-ultrasound scanner and UBM demonstrated the Intraocular Lens (IOL) was centered in the bag. No chafing in all directions was detected between IOL and iris/ciliary body. The proposed diagnoses were iridocyclitis and secondary glaucoma of the right eye. Interventions: The symptoms were not improved after antiinflammation and intraocular pressure (IOP) lowering treatment for 1 month. B-ultrasonography was applied in horizontal, sitting, and head-down positions. The results demonstrated movements of IOLs when position changed. The IOLs were in contact with the iris pigment epithelium in sitting position and head-down positions but not in horizontal position. The dynamic interactions between IOLs and iris/ciliary body implied a diagnosis of UGH syndrome. The IOLs were then extracted. Outcomes: Two weeks after the IOLs explantation, the IOP significantly reduced to a normal level in both eyes. Ten-month follow-up showed that the IOP was maintained at a normal level. Lessons: The chronically intermittent chafing between IOL and iris in specific head positions would also lead to UGH syndrome. Dynamic application of B-ultrasonography in various head positions could be useful in the diagnosis of an untypical UGH syndrome.

Purpose To evaluate the image resolutions in global teleophthalmology using 2 smartphones. Methods A standard resolution test card and two human eyes (one eye wearing a contact lens) were imaged with an iPhone 4S adapted for slitlamp biomicroscopy in Hangzhou, Zhejiang, China. An iPhone 5 was used as a receptor on the other side of the world (Miami, FL). The real-time images and still images were transferred through Skype from Hangzhou to Miami during slitlamp eye examination. The real-time resolutions captured and displayed on the both phones at different places were measured. There was vocal communication concurrent between two parties during image transference and the examination by the slitlamp biomicroscopy performed in Hangzhou. Results The real-time and still images were transferred between two cities while two operators talked with each other smoothly and without difficulty. The viewer in Miami was able to instruct the operator in Hangzhou to thoroughly examine the eye using the slitlamp microscopy with different magnification settings and illumination settings. The resolution of the still images recorded in the iPhone in Hangzhou was higher than that of the real-time images on the screen. The main features of the eye were recognizable in real-time. Conclusions The study demonstrated that global teleophthalmology is feasible using two smartphones. The system is simple, portable, and affordable, and the image quality in still and real-time images is acceptable for real-time teleophthalmology.

To evaluate the resolutions of slit-lamp microscopy photography using various cameras.Evaluation of diagnostic test or technology.Healthy subjects were imaged with these adapted cameras through slit-lamp microscopy.A total of 8 cameras, including 6 custom-mounted slit-lamp cameras and 2 commercial slit-lamp cameras, were tested with standard slit-lamp microscopy devices for imaging of the eye. Various magnifications were used during imaging. A standard resolution test plate was used to test the resolutions at different magnifications. These outcomes were compared with commercial slit-lamp cameras.The main measurements included the display spatial resolutions, image spatial resolutions, and ocular resolutions. The outcome also includes the relationships between resolution and the pixel density of the displays and images.All cameras were successfully adapted to the slit-lamp microscopy, and high-quality ocular images were obtained. Differences in the display spatial resolutions were found among cameras [analysis of variance (ANOVA), P0.05]. Higher display resolutions were found with cameras using the high-definition multimedia interface (HDMI) compared with others, including cameras in smart phones. The display resolutions of smart phone displays were greater than cameras with video graphics array displays. The display spatial resolutions were found as a function of display pixel density (r0.95, P0.05) and magnification (r0.85, P0.05). Different image spatial resolutions were found among cameras (ANOVA, P0.05) as a function of image pixel density (r0.98, P0.05) and magnification (r0.85, P0.05). The commercial slitlamp with a single lens reflex camera yielded the highest image spatial resolution. However, the ocular resolution through binocular viewing of the slit-lamp microscopy was found to have the highest resolution compared with the display and image spatial resolutions of all of the cameras.Several cameras can be adapted with slit-lamp microscopy for ophthalmic imaging, yielding various display and image spatial resolutions. However, the resolution seemed to not be as good as ocular viewing through the slit-lamp biomicroscope.

The total development trend in the world is towards continuously lower of sulfur content as a quality standard of diesel fuels. Integrating of an oxidative desulfurization unit with a conventional hydrotreating unit can bring benefits to producing low and/or ultra-low sulfur diesel fuels. Using the hydrotreated Middle East diesel fuel as a feedstock, four processes of the oxidative desulfurization have been studied: a hydrogen peroxide–acetic acid system and a Fenton's reagent system both without/with ultrasound. Results showed that the oxidative desulfurization reaction mechanics fitted apparent first-order kinetics. The addition of Fenton's reagent could enhance the oxidative desulfurization efficiency for diesel fuels and sono-oxidation treatment in combination with Fenton's reagent shows a good synergistic effect. Under our best operating condition for the oxidative desulfurization: temperature 313 K, ultrasonic power 200 W, ultrasonic frequency 28 kHz, Fe 2+ /H 2 O 2 0.05 mol/mol, pH 2.10 in aqueous phase and reaction time 15 min, the sulfur content in the diesel fuels was decreased from 568.75 μg/g to 9.50 μg/g.

Limacon gear, which is put forward as a new non-circular gear recent year, is used in the various equpiments. In order to analyse the transmission characteristics of limacon gear, three-dimensional models of one-order and high-order limacon gear were established according to its basic meshing theory, and imported the models into ADAMS through the interface, simulation analysis of its motion state in the actual working condition was done, angular velocity curve of conjugate gear and contact forces were obtained. By the comparative analysis on the theoretical and emulational angular velocity curve, relative error of maximum and minimum angular velocity was obtained and the reasons were analyzed. The simulation results further verify the correctness of the modeling method and the theoretical transmission characteristics of limacon gear pair.

To investigate changes of whole eye axial biometry during accommodation using ultra-long scan depth optical coherence tomography (UL-OCT).Prospective, observational case series.Twenty-one adult subjects were enrolled. Using UL-OCT, the left eye of each subject was imaged with relaxed diopters (0 D) and accommodative stimuli (+6 D). Full eye biometry included central corneal thickness (CCT), anterior chamber depth (ACD), lens thickness, vitreous length, and axial length (AL).During accommodation (+6 D), the axial biometry of the whole eye changed significantly. Compared to the rest state, ACD at the accommodative state decreased significantly from 3.128 ± 0.305 mm to 2.961 ± 0.298 mm (paired t test, P.001). The lens thickness increased significantly from 3.723 ± 0.237 mm to 3.963 ± 0.234 mm (P.001). The vitreous length decreased significantly from 17.129 ± 0.864 mm to 17.057 ± 0.848 mm (P.001). AL was 24.519 ± 0.917 mm at the rest state and increased to 24.545 ± 0.915 mm with +6 D accommodation stimulus. The elongated AL of 26.1 ± 13.4 μm between the rest and accommodative states was significant (P.001).During accommodation, whole eye axial biometry changed, including a decrease in ACD and vitreous length and an increase in lens thickness and AL. UL-OCT provides an alternative method that is suitable for full eye biometry during accommodation.

To assess the repeatability of axial biometry of the entire eye using ultra-long scan depth optical coherence tomography (OCT) and to investigate the agreement with IOLMaster measurements (Carl Zeiss Meditec).Prospective, observational case series.There were 37 adult subjects enrolled in group 1 and 12 adult subjects enrolled in group 2. Using ultra-long scan depth OCT, the left eyes of these groups were measured in 2 separate sessions. The images were processed by a manual method and custom-developed automatic software. A model eye was imaged for verification. The subjects in group 2 were imaged using ultra-long scan depth OCT and using the IOLMaster for axial length measurement comparison.All measured parameters of the model eye matched the geometric parameters. In group 1, there were no significant differences in all measured parameters using automatic and manual segmentation methods (P.05, paired t test). The percentage of coefficient of repeatability of segments ranged from 0.3% to 3.9%. The corresponding interclass correlation coefficients ranged from 0.946 to 0.999. The correlation between the results using automatic and manual segmentation methods appeared to be strong (R(2) = 0.999; P.05). In group 2, the axial length of the eye measured by the IOLMaster matched the results obtained by ultra-long scan depth OCT with the automatic method (R1(2) = 0.987; P.05) and the manual method (R2(2) = 0.988; P.05).Automatic axial biometry using ultra-long scan depth OCT successfully measured each segment of the entire eye with good repeatability. With further development of automatic segmentation, ultra-long scan depth OCT seems to be a promising tool in the axial biometry of the entire eye.