Missing Recovery: Single Image Reflection Removal based on Auxiliary Prior Learning.

Photographs taken through a glass window are susceptible to disturbances due to reflection. Therefore, single image reflection removal is crucial to image quality enhancement. In this paper, a novel learning architecture that can address this ill-posed problem is proposed. First, a novel reflection removal pipeline was designed to reconstruct the missing information caused by the camera imaging process using the proposed missing recovery network. Second, to address the issues in existing reflection removal strategies, we revisit several auxiliary priors and integrate them by defining an energy function. To solve the energy function, a convolutional neural network-based optimization scheme was proposed. Finally, we investigated the dark channel responses of reflection and clean images and found an interesting way to distinguish between these two types of images. We prove this property mathematically and propose a novel loss function called dark channel loss to improve performance. Experiments show that the proposed method outperforms state-of-the-art reflection removal methods both quantitatively and qualitatively.

Spectral-Temporal Receptive Field-Based Descriptors and Hierarchical Cascade Deep Belief Network for Guitar Playing Technique Classification.

Music information retrieval is of great interest in audio signal processing. However, relatively little attention has been paid to the playing techniques of musical instruments. This work proposes an automatic system for classifying guitar playing techniques (GPTs). Automatic classification for GPTs is challenging because some playing techniques differ only slightly from others. This work presents a new framework for GPT classification: it uses a new feature extraction method based on spectral-temporal receptive fields (STRFs) to extract features from guitar sounds. This work applies a supervised deep learning approach to classify GPTs. Specifically, a new deep learning model, called the hierarchical cascade deep belief network (HCDBN), is proposed to perform automatic GPT classification. Several simulations were performed and the datasets of: 1) data on onsets of signals; 2) complete audio signals; and 3) audio signals in a real-world environment are adopted to compare the performance. The proposed system improves upon the F-score by approximately 11.47% in setup 1) and yields an F-score of 96.82% in setup 2). The results in setup 3) demonstrate that the proposed system also works well in a real-world environment. These results show that the proposed system is robust and has very high accuracy in automatic GPT classification.

Impact of blood pressure control on retinal microvasculature in patients with chronic kidney disease.

Chronic kidney disease (CKD) is an emerging disease worldwide. We investigated the relationship between blood pressure (BP) control and parafoveal retinal microvascular changes in patients with CKD. This case-control study enrolled 256 patients with CKD (stage 3-5) and 70 age-matched healthy controls. Optical coherence tomography angiography showed lower superficial vascular plexus (SVP) vessel density, lower deep vascular plexus (DVP) vessel density, and larger SVP flow void area in the CKD group. The BP parameters at enrollment and during the year before enrollment were collected in patients with CKD. Partial correlation was used to determine the relationship between BP parameters and microvascular parameters after controlling for age, sex, diabetes mellitus, axial length, and intraocular pressure. The maximum systolic blood pressure (SBP) (p = 0.003) and within-patient standard deviation (SD) of SBP (p = 0.006) in 1 year were negatively correlated with SVP vessel density. The average SBP (p = 0.040), maximum SBP (p = 0.001), within-patient SD of SBP (p < 0.001) and proportion of high BP measurement (p = 0.011) in 1 year were positively correlated with the SVP flow void area. We concluded that long-term SBP was correlated with SVP microvascular injury in patients with CKD. Superficial retinal microvascular changes may be a potential biomarker for prior long-term BP control in these patients.

Reweighting of the sensory inputs for postural control in patients with cervical spondylotic myelopathy after surgery.

BACKGROUND: Cervical spondylotic myelopathy (CSM) is a degenerative cervical disease in which the spinal cord is compressed. Patients with CSM experience balance disturbance because of impaired proprioception. The weighting of the sensory inputs for postural control in patients with CSM is unclear. Therefore, this study investigated the weighting of sensory systems in patients with CSM. METHOD: Twenty-four individuals with CSM (CSM group) and 24 age-matched healthy adults (healthy control group) were analyzed in this observational study. The functional outcomes (modified Japanese Orthopaedic Association Scale [mJOA], Japanese Orthopaedic Association Cervical Myelopathy Questionnaire [JOACMEQ], Nurick scale) and static balance (eyes-open and eyes-closed conditions) were assessed for individuals with CSM before surgery, 3 and 6 months after surgery. Time-domain and time-frequency-domain variables of the center of pressure (COP) were analyzed to examine the weighting of the sensory systems. RESULTS: In the CSM group, lower extremity function of mJOA and Nurick scale significantly improved 3 and 6 months after surgery. Before surgery, the COP mean velocity and total energy were significantly higher in the CSM group than in the control group for both vision conditions. Compared with the control group, the CSM group exhibited lower energy content in the moderate-frequency band (i.e., proprioception) and higher energy content in the low-frequency band (i.e., cerebellar, vestibular, and visual systems) under the eyes-open condition. The COP mean velocity of the CSM group significantly decreased 3 months after surgery. The energy content in the low-frequency band (i.e., visual and vestibular systems) of the CSM group was closed to that of the control group 6 months after surgery under the eyes-open condition. CONCLUSION: Before surgery, the patients with CSM may have had compensatory sensory weighting for postural control, with decreased weighting on proprioception and increased weighting on the other three sensory inputs. After surgery, the postural control of the patients with CSM improved, with decreased compensation for the proprioceptive system from the visual and vestibular inputs. However, the improvement remained insufficient because the patients with CSM still had lower weighting on proprioception than the healthy adults did. Therefore, patients with CSM may require balance training and posture education after surgery. TRIAL REGISTRATION: Trial Registration number: NCT03396055 Name of the registry: ClinicalTrials.gov Date of registration: January 10, 2018 - Retrospectively registered Date of enrolment of the first participant to the trial: October 19, 2015.

A telesurveillance system with automatic electrocardiogram interpretation based on support vector machine and rule-based processing.

BACKGROUND: Telehealth care is a global trend affecting clinical practice around the world. To mitigate the workload of health professionals and provide ubiquitous health care, a comprehensive surveillance system with value-added services based on information technologies must be established. OBJECTIVE: We conducted this study to describe our proposed telesurveillance system designed for monitoring and classifying electrocardiogram (ECG) signals and to evaluate the performance of ECG classification. METHODS: We established a telesurveillance system with an automatic ECG interpretation mechanism. The system included: (1) automatic ECG signal transmission via telecommunication, (2) ECG signal processing, including noise elimination, peak estimation, and feature extraction, (3) automatic ECG interpretation based on the support vector machine (SVM) classifier and rule-based processing, and (4) display of ECG signals and their analyzed results. We analyzed 213,420 ECG signals that were diagnosed by cardiologists as the gold standard to verify the classification performance. RESULTS: In the clinical ECG database from the Telehealth Center of the National Taiwan University Hospital (NTUH), the experimental results showed that the ECG classifier yielded a specificity value of 96.66% for normal rhythm detection, a sensitivity value of 98.50% for disease recognition, and an accuracy value of 81.17% for noise detection. For the detection performance of specific diseases, the recognition model mainly generated sensitivity values of 92.70% for atrial fibrillation, 89.10% for pacemaker rhythm, 88.60% for atrial premature contraction, 72.98% for T-wave inversion, 62.21% for atrial flutter, and 62.57% for first-degree atrioventricular block. CONCLUSIONS: Through connected telehealth care devices, the telesurveillance system, and the automatic ECG interpretation system, this mechanism was intentionally designed for continuous decision-making support and is reliable enough to reduce the need for face-to-face diagnosis. With this value-added service, the system could widely assist physicians and other health professionals with decision making in clinical practice. The system will be very helpful for the patient who suffers from cardiac disease, but for whom it is inconvenient to go to the hospital very often.

Two-dimensional orthogonal DCT expansion in trapezoid and triangular blocks and modified JPEG image compression.

In the conventional JPEG algorithm, an image is divided into eight by eight blocks and then the 2-D DCT is applied to encode each block. In this paper, we find that, in addition to rectangular blocks, the 2-D DCT is also orthogonal in the trapezoid and triangular blocks. Therefore, instead of eight by eight blocks, we can generalize the JPEG algorithm and divide an image into trapezoid and triangular blocks according to the shapes of objects and achieve higher compression ratio. Compared with the existing shape adaptive compression algorithms, as we do not try to match the shape of each object exactly, the number of bytes used for encoding the edges can be less and the error caused from the high frequency component at the boundary can be avoided. The simulations show that, when the bit rate is fixed, our proposed algorithm can achieve higher PSNR than the JPEG algorithm and other shape adaptive algorithms. Furthermore, in addition to the 2-D DCT, we can also use our proposed method to generate the 2-D complete and orthogonal sine basis, Hartley basis, Walsh basis, and discrete polynomial basis in a trapezoid or a triangular block.

Salient region detection improved by principle component analysis and boundary information.

Salient region detection is useful for several image-processing applications, such as adaptive compression, object recognition, image retrieval, filter design, and image retargeting. A novel method to determine the salient regions of images is proposed in this paper. The L0 smoothing filter and principle component analysis (PCA) play important roles in our framework. The L0 filter is extremely helpful in characterizing fundamental image constituents, i.e., salient edges, and can simultaneously diminish insignificant details, thus producing more accurate boundary information for background merging and boundary scoring. PCA can reduce computational complexity as well as attenuate noise and translation errors. A local-global contrast is then used to calculate the distinction. Finally, image segmentation is used to achieve full-resolution saliency maps. The proposed method is compared with other state-of-the-art saliency detection methods and shown to yield higher precision-recall rates and F-measures.

Improved implementation algorithms of the two-dimensional nonseparable linear canonical transform.

The two-dimensional nonseparable linear canonical transform (2D NSLCT), which is a generalization of the fractional Fourier transform and the linear canonical transform, is useful for analyzing optical systems. However, since the 2D NSLCT has 16 parameters and is very complicated, it is a great challenge to implement it in an efficient way. In this paper, we improved the previous work and propose an efficient way to implement the 2D NSLCT. The proposed algorithm can minimize the numerical error arising from interpolation operations and requires fewer chirp multiplications. The simulation results show that, compared with the existing algorithm, the proposed algorithms can implement the 2D NSLCT more accurately and the required computation time is also less.

Reversible integer color transform.

In this correspondence, we introduce a systematic algorithm that can convert any 3 x 3 color transform into a reversible integer-to-integer transform. We also discuss the ways to improve accuracy and reduce implementation complexity. We derive the integer RGB-to-KLA, IV1 V2, YCbCr, DCT, YUV, and YIQ transforms that are optimal in accuracy.

Generalized prolate spheroidal wave functions for optical finite fractional Fourier and linear canonical transforms.

Prolate spheroidal wave functions (PSWFs) are known to be useful for analyzing the properties of the finite-extension Fourier transform (fi-FT). We extend the theory of PSWFs for the finite-extension fractional Fourier transform, the finite-extension linear canonical transform, and the finite-extension offset linear canonical transform. These finite transforms are more flexible than the fi-FT and can model much more generalized optical systems. We also illustrate how to use the generalized prolate spheroidal functions we derive to analyze the energy-preservation ratio, the self-imaging phenomenon, and the resonance phenomenon of the finite-sized one-stage or multiple-stage optical systems.

Eigenfunctions of the offset Fourier, fractional Fourier, and linear canonical transforms.

The offset Fourier transform (offset FT), offset fractional Fourier transform (offset FRFT), and offset linear canonical transform (offset LCT) are the space-shifted and frequency-modulated versions of the original transforms. They are more general and flexible than the original ones. We derive the eigenfunctions and the eigenvalues of the offset FT, FRFT, and LCT. We can use their eigenfunctions to analyze the self-imaging phenomena of the optical system with free spaces and the media with the transfer function exp[j(h2x2 + h1x + h0)] (such as lenses and shifted lenses). Their eigenfunctions are also useful for resonance phenomena analysis, fractal theory development, and phase retrieval.