Response Filter Research Articles

Poster Session: Study of contrast detection filters for images.

In perceptual contrast imaging, which removes only the contrast information humans cannot perceive from images, detecting local contrast per frequency is essential. It is known that simple and complex cells in the primary visual cortex (V1) are involved in human contrast detection. Therefore, in this study, we experimented with filters modeled after simple and complex cells for contrast detection to determine which filter is more suitable for perceptual contrast imaging. In the experiment, we modeled simple cells with Gabor filters and complex cells with the energy model, which is the square root of the sum of the squares of two Gabor filter responses with a phase difference of π/2. We used each model for contrast detection to generate perceptual contrast images and conducted subjective evaluation experiments to assess the distinguishability between the produced perceptual contrast images and the input images. In the experiment, subjects memorized an original input image to be selected in the next step. Then, subjects selected an image they believed to be the original in a two-interval forced-choice procedure. The results show that the accuracy in the complex cell model experiments was significantly close to 0.5, indicating no visual differences between the perceptual contrast images and their original, while the experiments of the simple cell model were not. Therefore, we concluded the complex cell model is more suitable for contrast detection in perceptual contrast imaging.

Novel Multiple-Input Single-Output Shadow Filter with Improved Passband Gain Using Multiple-Input Multiple-Output DDTAs

This paper presents a multiple-input single-output (MISO) shadow filter implemented using multiple-input differential difference transconductance amplifiers (MI-DDTAs). The MI-DDTA’s multiple inputs are realized through the multiple-input bulk-driven MOS transistor (MI-BD MOST) technique. Leveraging the multiple-input capability of the DDTA, various filter responses—low-pass filter (LPF), high-pass filter (HPF), band-pass filter (BPF), band-stop filter (BSF), and all-pass filter (APF)—can be efficiently achieved by appropriately configuring the input signals. The natural frequency and quality factor of the shadow filter can be independently tuned using external amplifiers. Unlike conventional shadow filters, where adjusting the quality factor or natural frequency impacts the passband gain, this design ensures a constant unity passband gain. The MI-DDTA operates at a supply voltage of 0.5 V and consumes 385.8 nW of power for setting current Iset = 14 nA. The proposed MI-DDTA and shadow filter are designed and validated through simulations in the Cadence design environment, using a 0.18 µm CMOS process provided by TSMC (Taiwan Semiconductor Manufacturing Company Limited).

Simultaneous Regularity Contrast and Luminance Polarity.

Texture regularity, for example, the repeating pattern of a carpet, brickwork, or tree bark, is a ubiquitous feature of the visual world. The perception of regularity has generally been studied using multi-element textures whose regularity is manipulated by the addition of random jitter to the elements' nominal positions. Here, we investigate the selectivity of regularity perception for the luminance contrast polarities of the elements. Our psychophysical tool was simultaneous regularity contrast, or SRC, the phenomenon in which the perceived regularity of a central test texture is shifted away from that of the surrounding regularity. Stimuli were composed of arrays of dark and/or white Gaussian elements. Surround and center test textures consisted of either the same ("congruent") or opposite ("incongruent") polarities. In addition, we tested a "mixed" condition consisting of a random mixture of polarities in both the surround and test. The perceived regularity of the test was measured using a match stimulus with the same polarity dimension as the test. The regularity of the match stimulus was adjusted on each trial using a forced-choice staircase procedure and the point-of-subjective equality between the match and test regularities was estimated from the resulting psychometric functions. SRC was observed in both congruent and incongruent conditions, but with the mixed condition, the perceived regularity of the test was shifted toward rather than away from the surround regularity, an example of assimilation, not contrast. The analysis revealed no significant difference in the magnitude of SRC between the congruent and incongruent conditions, suggesting that SRC could be mediated solely by polarity agnostic mechanisms, although there are other possible explanations for the "null" result. However, trend analysis using a non-linear (sigmoidal-shaped) function indicated a significant difference between the congruent and incongruent conditions, which, together with the mixed polarity results, suggests the presence of at least some polarity selective mechanisms. Previous reports have suggested that regularity perception is encoded by the "peakedness" in the distribution of spatial-frequency-tuned linear filter responses. We modelled SRC quantitatively by incorporating peakedness with spatial-frequency-selective surround inhibition and found that the model gave a good account of the SRC data. Possible reasons for the assimilation effect-with the mixed polarity condition are discussed.

Maximum Correntropy Criterion Kalman/Allan Variance-Assisted FIR Integrated Filter for Indoor Localization.

To obtain more accurate information on using an inertial navigation system (INS)-based integrated localization system, an integrated filter with maximum correntropy criterion Kalman filter (mccKF) and finite impulse response (FIR) is proposed for the fusion of INS-based multisource sensor data in this work. In the realm of medical applications, precise localization is crucial for various aspects, such as tracking the movement of a medical instrument within the human body or monitoring its position in the human body during procedures. This study uses ultra-wideband (UWB) technology to rectify the position errors of the INS. In this method, the difference between the positions of the INS and UWB is used as the measurement of the filter. The main data fusion filter in this study is the mccKF, which utilizes the maximum correntropy criterion (mcc) method to enhance the robustness of the Kalman filter (KF). This filter is used for fusing data from multiple sources, including the INS. Moreover, we use the Mahalanobis distance to verify the performance of the mccKF. If the performance of the mccKF is lower than the preset threshold, the Allan Variance-assisted FIR filter is used to replace the mccKF, which is designed in this work. This adaptive approach ensures the resilience of the system in demanding medical environments. Two practical experiments were performed to evaluate the effectiveness of the proposed approach. The findings indicate that the mccKF/FIR integrated method reduces the localization error by approximately 32.43% and 37.5% compared with the KF and mccKF, respectively. These results highlight the effectiveness of the proposed approach.

Learning temporal regularized spatial-aware deep correlation filter tracking via adaptive channel selection.

Learning temporal regularized spatial-aware deep correlation filter tracking via adaptive channel selection.

Analog Gaussian-Shaped Filter Design and Current Mode Compensation for Dot-Matrix TSP Readout Systems

In-cell touch and display integrated panels, along with their integrated readout systems, are widely adopted in mobile devices for their cost-effectiveness and compact design. This paper proposes an analog Gaussian-shaped filter and a current mode compensation technique for dot-matrix Touch Screen Panel (TSP) readout systems. Specifically, this article presents a noise management strategy for both intrinsic and external noise, offering simulation guidelines for determining intrinsic circuit noise levels in relation to scan time and enhancing external noise immunity through the Gaussian-shaped filter response. The system achieved an intrinsic SNR of 66 dB with a 200 kHz TSP driving frequency and a 160 μs scan time, while the 4-bit quantized Gaussian coefficients filter provided 33 dB noise suppression for out-of-band noise. The compensation error in the dot-matrix capacitance compensation was measured at 1.24 pF, which corresponds to a 0.078% deviation. The simulated power consumption of the proposed readout system is 24 mW, with a layout area of 1.017 mm2 for the 10-channel readout front-end.

Content-based image retrieval using COSFIRE descriptors with application to radio astronomy

ABSTRACT The morphologies of astronomical sources are highly complex, making it essential not only to classify the identified sources into their predefined categories but also to determine the sources that are most similar to a given query source. Image-based retrieval is essential, as it allows an astronomer with a source under study to ask a computer to sift through the large archived database of sources to find the most similar ones. This is of particular interest if the source under study does not fall into a ‘known’ category (anomalous). Our work uses the trainable COSFIRE (Combination of Shifted Filter Responses) approach for image retrieval. COSFIRE filters are automatically configured to extract the hyperlocal geometric arrangements that uniquely describe the morphological characteristics of patterns of interest in a given image; in this case astronomical sources. This is achieved by automatically examining the shape properties of a given prototype source in an image, which ultimately determines the selectivity of a COSFIRE filter. We further utilize hashing techniques, which are efficient in terms of required computation and storage, enabling scalability in handling large data sets in the image retrieval process. We evaluated the effectiveness of our approach by conducting experiments on a benchmark data set of radio galaxies, containing 1180 training images and 404 test images. Notably, our approach achieved a mean average precision of 91 per cent for image retrieval, surpassing both DenseNet-161 and group-equivariant convolutional neural networks (G-CNNs). Moreover, our approach is significantly more computationally efficient compared to both DenseNet-161 and G-CNNs.

Design of dual-band filtering patch antenna slot coupling feed

Abstract A novel microstrip filtering antenna with slot coupling feed is presented in this work. An asymmetric interdigital coupling structure is used for the feed to excite the patch antenna through gap-fed coupling. Introducing a U-shaped slot on the patch surface modifies the current path to attain different resonant modes. The asymmetric coupled fingers in the low-frequency band generate a radiation null of −64 dBi, while additional resonances introduced in the high band broaden the bandwidth from 4.9 to 5.3 GHz. A horizontally shorted microstrip branch produces another null point of −28 dBi between the bands, enabling steeper roll-off and further improvement in frequency selectivity. The proposed filtering antenna provides a flexible filter response without requiring extra filtering circuits, with appreciable peak gains (6.1 dBi and 7.1 dBi) and stable radiation characteristics. This makes it suitable for WLAN (Wireless Local Area Network) and ISM (Industrial Scientific Medical) systems applications.

IncARMAG: A convolutional neural network with multi-level autoregressive moving average graph convolutional processing framework for medical image classification

IncARMAG: A convolutional neural network with multi-level autoregressive moving average graph convolutional processing framework for medical image classification

On-chip power efficient MHz to GHz tunable Brillouin microwave photonic filters

Stimulated Brillouin scattering (SBS) offers optically tunable, narrowband gain and loss resonances, which makes it ideal for optical filtering and signal processing applications. However, broadband filters utilizing Brillouin scattering typically necessitate multiple pump sources or linear frequency modulated waveforms, which necessitate the use of overall high pump powers to achieve broadband and multi-band filter configurations. To overcome the trade-off between bandwidth and power consumption, we introduce an on-chip broadband bandpass filter scheme leveraging the SBS response and the concept of radio frequency interference. We present the MHz to GHz reconfigurable microwave photonic filters on a chip. The bandwidth is tunable from 30 MHz to 1.2 GHz at a total pump power of only 15 dBm, corresponding to a figure of merit (maximum bandwidth/pump power) improvement of a factor of 25 compared to previously reported work. Furthermore, the filter exhibits frequency tuning capabilities ranging from 10 to 40 GHz (X to Ka-band) and can be configured from single- to multi-bandpass filter responses and switched from bandpass to bandstop.

Enhanced Vernier Effect in Cascaded Fiber Loop Interferometers for Improving Temperature Sensitivity

This work presents a high-sensitivity temperature sensing system utilizing an enhanced Vernier effect implemented in cascaded fiber loop interferometers. High-sensitivity temperature sensors based on the Vernier effect have broad application prospects, but the sensitivity of traditional measurement schemes is difficult to improve further due to the limited variation in the difference between two free spectrum ranges (FSRs). Our sensing system incorporates two fiber loop interferometers and a single-mode fiber to form a Vernier spectral response, characterized by two complementary optical filter responses. As the temperature of the sensing fiber changes, one FSR decreases, and the other increases, respectively, enhancing the difference value between the two FSRs to form an enhanced Vernier effect. Experimental results demonstrate that the temperature sensitivity of a traditional Vernier effect measurement is only −298.29 kHz/°C, while our proposed enhanced Vernier effect sensing system achieves a sensitivity of 618.14 kHz/°C, which is 92 times higher than that of a two-arm optical carrier-based microwave interferometry (OCMI) sensing system and 2.07 times higher than that of a traditional Vernier effect sensing system. This approach with an enhanced Vernier effect scheme based on cascaded fiber loop interferometers can be used to design high-sensitivity sensing systems for biometrics, smart cities, and the Internet of Things.

Low‐Loss Differential SPST Switch With Filtering Response and Common‐Mode Suppression

Low‐Loss Differential SPST Switch With Filtering Response and Common‐Mode Suppression

Denoising graph attention wavelet network: an interpretable multi-sensor information fusion diagnostic method for rotating systems

Abstract The strong noise and black-box nature of deep networks pose great challenges to the efficiency of utilizing multi-sensor data for fault diagnosis. To solve these issues, a denoising graph attention wavelet network (DGAWN) is proposed for multi-sensor information fusion fault diagnosis of rotating machinery. Considering the spatial relationship of multi-sensor measurement points, k-neighborhood graphs are first constructed to characterize the intrinsic association and topology of each sensor data. The graph attention mechanism is introduced into the DGAWN to enhance the feature mapping relationship of multi-sensor graph nodes. Importantly, a denoising graph wavelet convolution layer is developed by utilizing a graph wavelet operator followed by an adaptive thresholding denoising module, thus obtaining interpretable graph wavelet filter responses and enhancing noise immunity. With the synergy of the attention mechanism, the DGAWN model matches the interpretable features with fault attributes under noise environments by utilizing the locality and sparsity of the graph wavelet kernel. Finally, two challenging rotating machinery fault datasets are used to validate the proposed DGAWN method under noise environments. Compared with several state-of-the-art models, experimental results indicate that DGAWN achieves the highest average diagnostic accuracies of 99.9% and 100% and superior noise resistance.

Integrated optoelectronic oscillator frequency tuning based on a variable optical time delay line in a loop-back

An optoelectronic oscillator, a simple and cost-effective electro-optical system, is known for its ability to generate ultra-high-frequency signals with an ultra-low phase noise level. This system can produce high-frequency signals up to microwaves, often surpassing their electric analogs in these parameters. Some scientific groups have even developed optoelectronic oscillators based on integrated photonics to enhance size and mass characteristics. However, these generators do have their drawbacks, including long-term frequency instability and a significant frequency step (about tens of kilohertz). The first drawback is due to the high sensitivity of optoelectronic oscillators to environmental parameters, particularly ambient temperature. While the thermal stabilization problem of photonic integrated circuits has been addressed, stabilizing the output frequency and its control in integrated optoelectronic oscillators remains a relevant and ongoing challenge in the field. The paper explores the potential for controlling the output frequency of an optoelectronic oscillator by adjusting the optical delay line in the feedback loop. We introduce a scheme for an optoelectronic oscillator based on the silicon-on-insulator platform and provide a mathematical model for the oscillator with a tunable optical delay line in the feedback loop. The tuning step for the output frequency is 50 MHz/ps, calculated analytically. We conducted numerical simulations using Ansys Lumerical Interconnect software to verify the mathematical model. The results show that the output frequency can be tuned in the range of 4.28 GHz, with a frequency step ranging from 30 to 60 MHz/ps. The maximum frequency deviation from the simulation is 20 MHz, approximately 0.2% of the output frequency. The obtained results demonstrate the effectiveness of the proposed frequency control method for various applications of the optoelectronic oscillator. The linear mathematical model we have developed can be used to calculate the output frequency of the optoelectronic oscillator while accounting for nonlinear errors. However, creating a comprehensive mathematical model that considers the nonlinear effects of integrated photonics will require further elaboration and development in future studies. The frequency tuning approach we have presented can be implemented using discrete optical components. However, in this case, the minimum frequency tuning step is limited by the discrete time delay in the optical delay line. On the other hand, an integrated optical delay line offers similar time delay control. This time delay tuning provides additional flexibility for controlling the output frequency within the optoelectronic oscillator eigenfrequencies, which are constrained by the frequency response of the notch filter.

Setting the phase-frequency response of digital filters with finite impulse response regardless of the amplitude-frequency response

The article discusses the implementation of a bandpass digital filter with independent adjustment of the phase-frequency response. By adjusting the phase-frequency characteristic we mean changing its angle of inclination, including the inverted form of the phase-frequency characteristic, that is, a negative angle of inclination. This makes it possible to obtain not only a small range of changes in the phase of the output signal in the filter passband, but also different signs of the slope of the phase-frequency characteristic. Carrying out such settings using known methods of implementing digital filters is impossible, since it contradicts physics. The solution to this problem is possible using a two-channel circuit for converting the input signal at multiple frequencies. The problems of adjusting the phase-frequency characteristic of a digital filter, independent of the amplitude-frequency characteristic, arise when designing digital self-oscillating systems. To fulfill the phase balance condition, it is necessary to eliminate ambiguity between frequencies and phase shifts. The range of changes in the phase-frequency characteristics in the passband should be within 2π. These requirements are met for analog filters, however, for digital filters, phase wrap is determined by the order of the filter. This leads to the fact that within the filter's passband, the phase-frequency characteristic zero can occur many times. The article presents a structural and mathematical model of a digital filter with an adjustable phase-frequency response, and its simulation was carried out in the MATLAB - Simulink environment. During the operation of the model, the dependences of the signal at the output of the blocks were obtained. Included in the structure of the model and it is shown that the experimental results do not contradict the mathematical model. The simulation result showed the effectiveness of tuning the phase-frequency response. Examples of setting the slope angle of the phase-frequency characteristic for three cases are shown: a small slope angle of the phase-frequency characteristic, zero equivalent phase-frequency characteristic and inverted (atypical) phase-frequency characteristic. Based on the results of the work, conclusions were drawn.

A Robust Visual Tracking Method based on Correlation Filter

Abstract Although the fast discriminative scale space tracking(fDSST) method shows superior performance for short time visual tracking, it is prone to tracking failure when the target is occluded or moving fast in case of long time tracking. To address this issue, we proposed a novel enhanced visual tracking method based on fDSST for robust tracking. Specifically, based on correlation filter response map we design a visual tracking status discrimination method by integrating Peak to Sidelobe Ratio(PSR) and the number of response peaks. Then, we design an adaptive model update method coupled with extended search area strategy to reduce the probability of target loss. Extensive experiments are performed on challenging benchmark sequences from Online Object Tracking Benchmark(OTB) with significant target occlusion and fast motion. Ours results show that the proposed approach improves the DP by 11.9% and AUC by 8.4% compared to the baseline fDSST, and operates at real-time.

Optimal Estimation of Local Motion-in-Depth with Naturalistic Stimuli.

Estimating the motion of objects in depth is important for behavior and is strongly supported by binocular visual cues. To understand both how the brain should estimate motion in depth and how natural constraints shape and limit performance in two local 3D motion tasks, we develop image-computable ideal observers from a large number of binocular video clips created from a dataset of natural images. The observers spatiotemporally filter the videos and nonlinearly decode 3D motion from the filter responses. The optimal filters and decoder are dictated by the task-relevant image statistics and are specific to each task. Multiple findings emerge. First, two distinct filter subpopulations are spontaneously learned for each task. For 3D speed estimation, filters emerge for processing either changing disparities over time or interocular velocity differences, cues that are used by humans. For 3D direction estimation, filters emerge for discriminating either left-right or toward-away motion. Second, the filter responses, conditioned on the latent variable, are well-described as jointly Gaussian, and the covariance of the filter responses carries the information about the task-relevant latent variable. Quadratic combination is thus necessary for optimal decoding, which can be implemented by biologically plausible neural computations. Finally, the ideal observer yields nonobvious-and in some cases counterintuitive-patterns of performance like those exhibited by humans. Important characteristics of human 3D motion processing and estimation may therefore result from optimal information processing in the early visual system.

Artificial Intelligence-Based DS-PSO Algorithm for Enhanced Frequency Response in Digital IIR Filters

Artificial Intelligence-Based DS-PSO Algorithm for Enhanced Frequency Response in Digital IIR Filters

Method for determining the effect of destabilizing factors of digital filters on the multicriteria synthesis of radio communications of aviation robotic devices

The purpose of the research is development of a method for determining the influence of destabilizing factors, such as the unevenness of the amplitude-frequency response (AFC) and the nonlinearity of the phase-frequency response (FFC), in non-recursive digital frequency selection filters on the target function of multicriteria structural-parametric synthesis of radio communications from the payload of aviation robotic devices.Methods. In conducting scientific research, probability theory, mathematical statistics, statistical radio engineering and computational mathematics were used. The study obtained mathematical expressions of the dependence of the probability of error per bit (BER) on the average signal-to-noise ratio when signals pass through symmetrical non-recursive digital filters with quadrature amplitude modulation with a positivity of no more than 4096, signal constellation irregularities of no more than 5, which are the basis for the energy balance of radio lines with aviation robotic devices. The expressions obtained are one of the "foundations" for the implementation of a multi-criteria synthesis of radio systems adaptive to a cognitively formed flight task. In this paper, all the results of the probability of reliability of the received data by means of communication are reflected for the case of rigid decision-making (Gray coding).Results. In the course of the conducted research: 1) a mathematical model of the structural-parametric synthesis of radio communications with aviation robotic devices is demonstrated, which is based on the criteria for minimizing: electronic computing resources with a given noise immunity; radiation outside the operating frequency band; peak factor signals; the effects of narrowband, broadband and structural interference in the radio channel at a given bandwidth; the use of structural and functional nodes specific to the existing information and communication environment; 2) the analytical dependence of the influence of the parameters of one of several structural and functional nodes - digital filters of frequency selection, on the criteria for the synthesis of communication means has been revealed. The paper considers the most common approximating functions - Kaiser, Tukey, Barlett-Khan, Chebyshev, and also shows the influence of the filter order on the equivalent of energy loss in information communication lines of aviation robotic devices.Conclusion. The scientific article presents a method for determining the effect of the uneven frequency response of a symmetrical non-recursive digital filter in devices for receiving and processing radio signals of multicriteria means of communication with the payload of aviation robotic devices. The proposed approach increases the reliability of determining the values of criteria for minimizing electronic computing resources for a given noise immunity, radiation outside the operating frequency band, as well as the use of structural and functional nodes characteristic of the existing information and communication environment.

Linear prediction coefficients correction method for digital speech processing systems with data compression based on the autoregressive model of a voice signal

The problem of distortion of the autoregressive model of the voice signal under the influence of additive background noise in digital speech processing systems with data compression based on linear prediction is considered. In the frequency domain, these distortions are observed in the weakening of the main formants responsible for the intelligibility of the speaker’s speech. To compensate for formant attenuation, it is proposed to modify the parameters of the autoregressive model (linear prediction coefficients) using the impulse response of a recursive shaping filter. Along with the amplitude amplification of the formants, their frequencies remain unchanged to make the speaker’s voice recognizable. The effectiveness of the method was studied experimentally using specially developed software. Based on the experimental results, conclusions were drawn about a significant increase in the relative level of formants in the power spectrum of the corrected voice signal.