In this paper, we present a new technique for mammogram enhancement using fast dyadic wavelet transform (FDyWT) based on lifted spline dyadic wavelets and normalized Tsallis entropy. First, a mammogram image is decom- posed into a multiscale hierarchy of low-subband and high-subband images using FDyWT. Then noise is suppressed using normalized Tsallis entropy of the local variance of the modulus of oriented high-subband images. After that, the wavelet coefficients of high-subbands are modified using a non-linear operator and finally the low-subband image at the first scale is modified with power law transformation to suppress background. Though FDyWT is shift-invariant and has better poten- tial for detecting singularities like edges, its performance depends on the choice of dyadic wavclcts. On the other hand, the nulnber of vanishing moments is an important characteristic of dyadic wavelets for singularity analysis because it provides an upper bound measurement for singularity characterization. Using lifting dyadic schemes, we construct lifted spline dyadic wavelets of different degrees with increased number of vanishing moments. We also examine the effect of these wavelets on mammogram enhancement. The method is tested on mammogram images, taken from MIAS (Mammographic Image Analysis Society) database, having various background tissue types and containing different abnormalities. The comparison with tile state-of-the-art contrast enhancement methods reveals that the proposed method performs better and the difference is statistically significant. 相似文献
Bearings play a crucial role in rotational machines and their failure is one of the foremost causes of breakdowns in rotary machinery. Their functionality is directly relevant to the operational performance, service life and efficiency of these machines. Therefore, bearing fault identification is very significant. The accuracy of fault or anomaly detection by the current techniques is not adequate. We propose a data mining-based framework for fault identification and anomaly detection from machine vibration data. In this framework, to capture the useful knowledge from the vibration data stream (VDS), we first pre-process the data using Fast Fourier Transform (FFT) to extract the frequency signature and then build a compact tree called SAFP-tree (sliding window associated frequency pattern tree), and propose a mining algorithm called SAFP. Our SAFP algorithm can mine associated frequency patterns (i.e., fault frequency signatures) in the current window of VDS and use them to identify faults in the bearing data. Finally, SAFP is further enhanced to SAFP-AD for anomaly detection by determining the normal behavior measure (NBM) from the extracted frequency patterns. The results show that our technique is very efficient in identifying faults and detecting anomalies over VDS and can be used for remote machine health diagnosis. 相似文献
Emotion recognition from speech signals is an interesting research with several applications like smart healthcare, autonomous voice response systems, assessing situational seriousness by caller affective state analysis in emergency centers, and other smart affective services. In this paper, we present a study of speech emotion recognition based on the features extracted from spectrograms using a deep convolutional neural network (CNN) with rectangular kernels. Typically, CNNs have square shaped kernels and pooling operators at various layers, which are suited for 2D image data. However, in case of spectrograms, the information is encoded in a slightly different manner. Time is represented along the x-axis and y-axis shows frequency of the speech signal, whereas, the amplitude is indicated by the intensity value in the spectrogram at a particular position. To analyze speech through spectrograms, we propose rectangular kernels of varying shapes and sizes, along with max pooling in rectangular neighborhoods, to extract discriminative features. The proposed scheme effectively learns discriminative features from speech spectrograms and performs better than many state-of-the-art techniques when evaluated its performance on Emo-DB and Korean speech dataset.
Parasitic absorption in transparent electrodes is one of the main roadblocks to enabling power conversion efficiencies (PCEs) for perovskite‐based tandem solar cells beyond 30%. To reduce such losses and maximize light coupling, the broadband transparency of such electrodes should be improved, especially at the front of the device. Here, the excellent properties of Zr‐doped indium oxide (IZRO) transparent electrodes for such applications, with improved near‐infrared (NIR) response, compared to conventional tin‐doped indium oxide (ITO) electrodes, are shown. Optimized IZRO films feature a very high electron mobility (up to ≈77 cm2 V?1 s?1), enabling highly infrared transparent films with a very low sheet resistance (≈18 Ω □?1 for annealed 100 nm films). For devices, this translates in a parasitic absorption of only ≈5% for IZRO within the solar spectrum (250–2500 nm range), to be compared with ≈10% for commercial ITO. Fundamentally, it is found that the high conductivity of annealed IZRO films is directly linked to promoted crystallinity of the indium oxide (In2O3) films due to Zr‐doping. Overall, on a four‐terminal perovskite/silicon tandem device level, an absolute 3.5 mA cm?2 short‐circuit current improvement in silicon bottom cells is obtained by replacing commercial ITO electrodes with IZRO, resulting in improving the PCE from 23.3% to 26.2%. 相似文献
As vehicle complexity and road congestion increase, combined with the emergence of electric vehicles, the need for intelligent transportation systems to improve on-road safety and transportation efficiency using vehicular networks has become essential. The evolution of high mobility wireless networks will provide improved support for connected vehicles through highly dynamic heterogeneous networks. Particularly, 5G deployment introduces new features and technologies that enable operators to capitalize on emerging infrastructure capabilities. Machine Learning (ML), a powerful methodology for adaptive and predictive system development, has emerged in both vehicular and conventional wireless networks. Adopting data-centric methods enables ML to address highly dynamic vehicular network issues faced by conventional solutions, such as traditional control loop design and optimization techniques. This article provides a short survey of ML applications in vehicular networks from the networking aspect. Research topics covered in this article include network control containing handover management and routing decision making, resource management, and energy efficiency in vehicular networks. The findings of this paper suggest more attention should be paid to network forming/deforming decision making. ML applications in vehicular networks should focus on researching multi-agent cooperated oriented methods and overall complexity reduction while utilizing enabling technologies, such as mobile edge computing for real-world deployment. Research datasets, simulation environment standardization, and method interpretability also require more research attention. 相似文献
Hierarchical core–shell (C–S) heterostructures composed of a NiO shell deposited onto stacked‐cup carbon nanotubes (SCCNTs) are synthesized by atomic layer deposition (ALD). A film of NiO particles (0.80–21.8 nm in thickness) is uniformly deposited onto the inner and outer walls of the SCCNTs. The electrical resistance of the samples is found to increase of many orders of magnitude with the increasing of the NiO thickness. The response of NiO–SCCNT sensors toward low concentrations of acetone and ethanol at 200 °C is studied. The sensing mechanism is based on the modulation of the hole‐accumulation region in the NiO shell layer upon chemisorption of the reducing gas molecules. The electrical conduction mechanism is further studied by the incorporation of an Al2O3 dielectric layer at NiO and SCCNT interfaces. The investigations on NiO–Al2O3–SCCNT, Al2O3–SCCNT, and NiO–SCCNT coaxial heterostructures reveal that the sensing mechanism is strictly related to the NiO shell layer. The remarkable performance of the NiO–SCCNT sensors toward acetone and ethanol benefits from the conformal coating by ALD, large surface area of the SCCNTs, and the optimized p‐NiO shell layer thickness followed by the radial modulation of the space‐charge region. 相似文献
Reducing transmit power is the most straightforward way towards more energy-efficient communications, but it results in lower SNRs at the receiver which can add a performance and/or complexity cost. At low SNRs, synchronization and channel estimation errors erode much of the gains achieved through powerful turbo and LDPC codes. Further expanding the turbo concept through an iterative receiver—which brings synchronization and equalization modules inside the loop—can help, but this solution is prohibitively complex and it is not clear what can and what cannot be a part of the iterative structure. This paper fills two important gaps in this field: (1) as compared to previous research which either focuses on a subset of the problem assuming perfect remaining parameters or is computationally too complex, we propose a proper partitioning of algorithm blocks in the iterative receiver for manageable delay and complexity, and (2) to the best of our knowledge, this is the first physical demonstration of an iterative receiver based on experimental radio hardware. We have found that for such a receiver to work, (1) iterative timing synchronization is impractical, iterative carrier synchronization can be avoided by using our proposed approach, while iterative channel estimation is essential, and (2) the SNR gains claimed in previous publications are validated in indoor channels. Finally, we propose a heuristic algorithm for simplifying the carrier phase synchronization in an iterative receiver such that computations of the log likelihood ratios of the parity bits can be avoided to strike a tradeoff between complexity and performance. 相似文献