Spatial Frequency Bands Research Articles

Learning a collaborative multiscale dictionary based on robust empirical mode decomposition

Abstract Dictionary learning is a challenging topic in many image processing areas. The basic goal is to learn a sparse representation from an overcomplete basis set. Due to combining the advantages of generic multiscale representations with the learning-based adaptivity, multiscale dictionary representation approaches have the power in capturing structural characteristics of natural images. However, the existing multiscale learning approaches still suffer from three main weaknesses: inadaptability to diverse scales of image data, sensitivity to noise and outliers, difficulty to determine the optimal dictionary structure. In this paper, we present a novel multiscale dictionary learning paradigm for sparse image representations based on an improved empirical mode decomposition. This powerful data-driven analysis tool for multi-dimensional signals can fully adaptively decompose the image into multiscale oscillating components according to intrinsic modes of data self. This treatment can obtain a robust and effective sparse representation, and meanwhile generates a raw dictionary at multiple geometric scales and spatial frequency bands. This dictionary is refined by selecting the optimal oscillating atom based on frequency clustering. In order to further enhance sparsity and generalization, a tolerant dictionary is learned using a coherence regularized model. A fast proximal scheme is developed to optimize this model. The multiscale dictionary is considered as the product of an oscillating dictionary and a tolerant dictionary. Experimental results demonstrate that the proposed method has superior performance compared with several competing methods for sparse image representations. We also have shown the promising results in image denoising application.

Influence of the pavement surface on the vibrations induced by heavy traffic in road bridges

The irregularity of the pavement surface governs the traffic-induced vibrations in road bridges, but it is either ignored or simulated by means of ideal pavements that differ significantly from real cases. This work presents a detailed dynamic analysis of a heavy truck crossing a 40 m span composite deck bridge using on-site measurements of different existing road profiles, as well as code-based ideal pavements. By activating or deactivating certain spatial frequency bands of the pavement, it is observed that the ranges 0.2–1 and 0.02–0.2 cycles/m are critical for the comfort of the pedestrians and the vehicle users, respectively. Well maintained roads with low values of the displacement power spectral density (PSD) associated with these spatial frequency ranges could reduce significantly the vibration on the sidewalks and, specially, in the vehicle cabin. Finally, a consistent road categorization for vibration assessment based on the PSD of the pavement irregularity evaluated at the dominant frequencies is proposed.

Cortical feedback signals generalise across different spatial frequencies of feedforward inputs

Visual processing in cortex relies on feedback projections contextualising feedforward information flow. Primary visual cortex (V1) has small receptive fields and processes feedforward information at a fine-grained spatial scale, whereas higher visual areas have larger, spatially invariant receptive fields. Therefore, feedback could provide coarse information about the global scene structure or alternatively recover fine-grained structure by targeting small receptive fields in V1. We tested if feedback signals generalise across different spatial frequencies of feedforward inputs, or if they are tuned to the spatial scale of the visual scene. Using a partial occlusion paradigm, functional magnetic resonance imaging (fMRI) and multivoxel pattern analysis (MVPA) we investigated whether feedback to V1 contains coarse or fine-grained information by manipulating the spatial frequency of the scene surround outside an occluded image portion. We show that feedback transmits both coarse and fine-grained information as it carries information about both low (LSF) and high spatial frequencies (HSF). Further, feedback signals containing LSF information are similar to feedback signals containing HSF information, even without a large overlap in spatial frequency bands of the HSF and LSF scenes. Lastly, we found that feedback carries similar information about the spatial frequency band across different scenes. We conclude that cortical feedback signals contain information which generalises across different spatial frequencies of feedforward inputs.

Object Categorization in Finer Levels Relies More on Higher Spatial Frequencies and Takes Longer.

The human visual system contains a hierarchical sequence of modules that take part in visual perception at different levels of abstraction, i.e., superordinate, basic, and subordinate levels. One important question is to identify the “entry” level at which the visual representation is commenced in the process of object recognition. For a long time, it was believed that the basic level had a temporal advantage over two others. This claim has been challenged recently. Here we used a series of psychophysics experiments, based on a rapid presentation paradigm, as well as two computational models, with bandpass filtered images of five object classes to study the processing order of the categorization levels. In these experiments, we investigated the type of visual information required for categorizing objects in each level by varying the spatial frequency bands of the input image. The results of our psychophysics experiments and computational models are consistent. They indicate that the different spatial frequency information had different effects on object categorization in each level. In the absence of high frequency information, subordinate and basic level categorization are performed less accurately, while the superordinate level is performed well. This means that low frequency information is sufficient for superordinate level, but not for the basic and subordinate levels. These finer levels rely more on high frequency information, which appears to take longer to be processed, leading to longer reaction times. Finally, to avoid the ceiling effect, we evaluated the robustness of the results by adding different amounts of noise to the input images and repeating the experiments. As expected, the categorization accuracy decreased and the reaction time increased significantly, but the trends were the same. This shows that our results are not due to a ceiling effect. The compatibility between our psychophysical and computational results suggests that the temporal advantage of the superordinate (resp. basic) level to basic (resp. subordinate) level is mainly due to the computational constraints (the visual system processes higher spatial frequencies more slowly, and categorization in finer levels depends more on these higher spatial frequencies).

Pyramid approach for the reduction of parallax-related artefacts in optical recordings of moving translucent volumes.

Functional optical imaging (OI) of intrinsic signals (like blood oxygenation coupled reflection changes) and of extrinsic properties of voltage sensitive probes (like voltage-sensitive dyes (VSD)) forms a group of invasive neuroimaging techniques, that possess up to date the highest temporal and spatial resolution on a meso- to macroscopic scale. There are different sources that contribute to the OI signal of which many are noise. In our previous works, we have used dense optical flow for the reduction of movement artefacts. The translucent surface of the cortex allows contributions from multiple depths. Due to the depth offield (DOF) effect, we get an implicit relation of depth and 2D frequency components. In this work, we introduce registration on the levels of a Laplacian pyramid to remove movement artefacts which have different motion components in different spatial frequency bands. This aims to resolve artefacts that remain after normal registration and are caused e.g. by parallax motion, dead pixels or dust on the sensor and other high frequent, moving particles on the cortex surface without the compromise of using high smoothness weights.

Developmental changes in the critical information used for facial expression processing

Facial expression recognition skills are known to improve across childhood and adolescence, but the mechanisms driving the development of these important social abilities remain unclear. This study investigates directly whether there are qualitative differences in child and adult processing strategies for these emotional stimuli. With a novel adaptation of the Bubbles reverse-correlation paradigm (Gosselin & Schyns, 2001), we added noise to expressive face stimuli and presented sub-sets of randomly sampled information from each image at different locations and spatial frequency bands across experimental trials. Results from our large developmental sample: 71 young children (6 –9years), 69 older children (10–13years) and 54 adults, uniquely reveal profiles of strategic information-use for categorisations of fear, sadness, happiness and anger at all ages. All three groups relied upon a distinct set of key facial features for each of these expressions, with fine-tuning of this diagnostic information (features and spatial frequency) observed across developmental time. Reported variability in the developmental trajectories for different emotional expressions is consistent with the notion of functional links between the refinement of information-use and processing ability.

Estimating Response Time in Spatial Frequency Band Experiments involving Visual Perception

The present study aimed to determine the optimum response time (RT) needed to identify images of everyday objects when filtered using different spatial frequency bands. Subjects were randomly presented with different images of familiar objects that were both serialized and progressive in their spatial frequencies. The time needed to recognize them was then measured. The results showed that the optimum RT for identifying an image filtered in different spatial frequency bands was approximately 2000 ms of exposure. Specifically, stimuli presented using spatial frequency bands with Gaussian filters of variance V26-V32, which were familiar and of medium size to the viewer, were recognized in a mean time of 2126 ms.

Full-band error control and crack-free surface fabrication techniques for ultra-precision fly cutting of large-aperture KDP crystals

Large-aperture potassium dihydrogen phosphate (KDP) crystals are widely used in the laser path of inertial confinement fusion (ICF) systems. The most common method of manufacturing half-meter KDP crystals is ultra-precision fly cutting. When processing KDP crystals by ultra-precision fly cutting, the dynamic characteristics of the fly cutting machine and fluctuations in the fly cutting environment are translated into surface errors at different spatial frequency bands. These machining errors should be suppressed effectively to guarantee that KDP crystals meet the full-band machining accuracy specified in the evaluation index. In this study, the anisotropic machinability of KDP crystals and the causes of typical surface errors in ultra-precision fly cutting of the material are investigated. The structures of the fly cutting machine and existing processing parameters are optimized to improve the machined surface quality. The findings are theoretically and practically important in the development of high-energy laser systems in China.

The influence of processing parameters on surface characteristics in micro-milling thin-walled slot on Elgiloy

High-leveled surface integrity which is significantly affected by processing parameters is required for machining military and aviation component. This paper firstly introduces the material (rolled Elgiloy), structure (with thin-walled slot), and machining method (micro-milling) of this component. To present surface structural traits, three-dimensional (3D) fractal dimension Ds based on variance method is proved to be a good index, compared with surface roughness. The micro-milled surface quality and structural characteristics are found to be different in various areas. The feed per tooth is revealed as the key factor that affects surface quality and complexity. Based on two-dimensional (2D) fractal method, trend of surfaces structural traits are evidently observed and found closely related to this processing method. These milled surfaces are then evaluated to be of good isotropy, considering the variance of slop and curvature of profiles. And surface anisotropy is found more sensitive to milling depth. With wavelet transform and 2D fractal method, key spatial frequency bands that affect anisotropy and complexity of the surface are revealed, respectively. The present analysis reveals the relationships between various areas of surfaces, anisotropy, and processing parameters in depth and helps to choose reasonable milling parameters for improving surface quality and complexity.

Non-uniform phase sensitivity in spatial frequency maps of the human visual cortex.

Just as a portrait painting can come from a collection of coarse and fine details, natural vision can be decomposed into coarse and fine components. Previous studies have shown that the early visual areas in the brain represent these components in a map-like fashion. Other studies have shown that these same visual areas can be sensitive to how coarse and fine features line up in space. We found that the brain actually jointly represents both the scale of the feature (fine, medium, or coarse) and the alignment of these features in space. The results suggest that the visual cortex has an optimized representation particularly for the alignment of fine details, which are crucial in understanding the visual scene. Complex natural scenes can be decomposed into their oriented spatial frequency (SF) and phase relationships, both of which are represented locally at the earliest stages of cortical visual processing. The SF preference map in the human cortex, obtained using synthetic stimuli, is orderly and correlates strongly with eccentricity. In addition, early visual areas show sensitivity to the phase information that describes the relationship between SFs and thereby dictates the structure of the image. Taken together, two possibilities arise for the joint representation of SF and phase: either the entirety of the cortical SF map is uniformly sensitive to phase, or a particular set of SFs is selectively phase sensitive - for example, greater phase sensitivity for higher SFs that define fine-scale edges in a complex scene. To test between these two possibilities, we constructed a novel continuous natural scene video whereby phase information was maintained in one SF band but scrambled elsewhere. By shifting the central frequency of the phase-aligned band in time, we mapped the phase-sensitive SF preference of the visual cortex. Using functional magnetic resonance imaging, we found that phase sensitivity in early visual areas is biased toward higher SFs. Compared to a SF map of the same scene obtained using linear-filtered stimuli, a much larger patch of areas V1 and V2 is sensitive to the phase alignment of higher SFs. The results of early areas cannot be explained by attention. Our results suggest non-uniform sensitivity to phase alignment in population-level SF representations, with phase alignment being particularly important for fine-scale edge representations of natural scenes.

Latency of evoked potentials in the tasks involving classification of images after wavelet filtration

We studied the characteristics of evoked potentials recorded during the recognition test based on four types of series of images subjected to the wavelet filtration: images of living objects containing either low frequency or high frequency portion of the spatial frequency spectrum, and imaging of non-living objects in the same two spatial frequency bands. Each subject had to classify the image either by its semantic feature (live-not live), or by its physical feature (low-high frequency). The purpose of this study was to compare the time characteristics of evoked potentials in these two types of tasks, which provides information on the time characteristics of categorization mechanisms of visual images. Analysis of the latent periods and amplitudes of the components of evoked potentials enabled to detect the occipital areas of the leads where the early components (up to 170 ms) are associated with spatial and frequency characteristics of the image, the frontal and temporal areas where the components of 170-200 ms correspond with the process of categorization, and the later frontal, central and parietal areas (300-500 ms) correspond with the process of error detection and the organization of motor response.

Efficient single pixel imaging in Fourier space

Single pixel imaging (SPI) is a novel technique capturing 2D images using a bucket detector with a high signal-to-noise ratio, wide spectrum range and low cost. Conventional SPI projects random illumination patterns to randomly and uniformly sample the entire scene’s information. Determined by Nyquist sampling theory, SPI needs either numerous projections or high computation cost to reconstruct the target scene, especially for high-resolution cases. To address this issue, we propose an efficient single pixel imaging technique (eSPI), which instead projects sinusoidal patterns for importance sampling of the target scene’s spatial spectrum in Fourier space. Specifically, utilizing the centrosymmetric conjugation and sparsity priors of natural images’ spatial spectra, eSPI sequentially projects two -phase-shifted sinusoidal patterns to obtain each Fourier coefficient in the most informative spatial frequency bands. eSPI can reduce requisite patterns by two orders of magnitude compared to conventional SPI, which helps a lot for fast and high-resolution SPI.

Glaucoma Increases Retinal Surface Contour Variability as Measured by Optical Coherence Tomography.

PurposeWe investigated the feasibility of glaucoma detection by measuring retinal surface contour variability (RSCV) using optical coherence tomography (OCT).MethodsThe peripapillary region in one eye of each participant was scanned over an 8 × 8 mm area with a swept source OCT prototype. The retinal surface contour was sampled at approximately 1.5- to 3.5-mm radius circles centered on the optic nerve head. The RSCV is defined as the average log value within a middle spatial frequency band of the Fourier transform to the elevation profile of the inner retinal surface. The spatial frequency band was optimized to distinguish glaucoma from normal. Nerve fiber layer thickness (NFLT) was sampled around a 1.7-mm radius circle. Glaucoma severity was assessed by automated static perimetry.ResultsWe enrolled 17 glaucomatous eyes and 17 healthy eyes. A great majority of the glaucoma group were in the early stage (visual field mean deviation average −2.48 ± 3.73 dB). Significant differences were found for RSCV between glaucoma and control eyes (P < 0.003) at all radii. The area under the receiver operating characteristic curve (AROC = 0.90) of RSCV was best at the 3.5-mm radius. This was not significantly better than NFLT (AROC = 0.84). With the 99% specificity, the glaucoma detection sensitivity was 53% for RSCV and 29% for NFLT (P = 0.13).ConclusionsRetinal surface contour variability was significantly increased in glaucoma patients. The diagnostic accuracy of RSCV was equal to NFLT in early glaucoma. Since the RSCV detects small-scale focal damage and the average NFLT measures global damage, they provide different diagnostic information that may be synergistic.

Synthesis of wave surface by frame loudspeaker array with limited spatial frequency band

We have studied a frame loudspeaker array for sound reproduction of three-dimensional television aiming at realizing clear localization and better sound depth control. Because such array has no secondary sources in its center region, the reproduced wave surface is sometimes deteriorated even if the least squares optimization will be used to reshape the surface. Over the past few decades, a considerable number of studies have been made on the sound field reproduction by loudspeaker array. Recently, higher order ambisonics (HOA) and near-field compensated HOA (NFC-HOA) attract attention of the researchers. It is well known that these methods can limit a spatial frequency band by controlling the degree of spherical function. If such limitation is introduced to the frame loudspeaker array, the deterioration of wave surface will be alleviated because the optimization will be easy to progress by matching only the low spatial frequency components of the reproduced wave surface with those of the reference. The co...

A New Wavelet Based Spatio-temporal Method for Magnification of Subtle Motions in Video

Video magnification is a computational procedure to reveal subtle variations during video frames that are invisible to the naked eye. A new spatio-temporal method which makes use of connectivity based mapping of the wavelet sub-bands is introduced here for exaggerating of small motions during video frames. In this method, firstly the wavelet transformed frames are mapped to connectivity space and then decomposed into different spatial frequency bands by applying Laplacian Pyramid to determine the pixels having more chance to be a part of a movement. Finally each candidate is partially magnified based on its time history. The performance of the proposed method is evaluated on real videos which contain several subtle motions. Parameters for performance evaluation are presented and obtained results are compared with one of state-of-the-art video magnification methods. Increased true positive rate in parallel with simultaneous decrease in false positive rate confirm the effectiveness of the proposed method in amplifying subtle motions.

The Effects of Letter Design Features and Aging on Legibility.

We examined how letter legibility is affected by letter design features (size, contrast, font-type, and complexity), and how this changes with age (by comparing younger [ N = 54] and older participants [ N = 54]). Types of letters in the study were Japanese characters, Latin alphabets, and digits. Participants verbally reported the letter presented on a Cathode Ray Tube (CRT) display in natural viewing conditions. The legible-contrast threshold was calculated as the contrast boundary of 50% correct answer. We found that letter complexity (i.e., number of strokes) did not affect legible contrast threshold, except for small (<0.35°) letters. Moreover, letter size even affected the threshold with a function similar but not identical to contrast sensitivity function. This supports the previous hypothesis proposed by Majaj et al. that optimal spatial frequency band exists for letter identification and shifts with its size. Another novel finding was the effect of aging on the legible-contrast threshold and how it related to other letter features. Constant aging effect on legible-contrast was found across letter sizes. Among older adults, the legible-contrast thresholds were 1.2 to 2 times those of younger adults. Moreover, the aging effects were greater for positive-contrast than for negative.

Increased spatial granularity of left brain activation and unique age/gender signatures: a 4D frequency domain approach to cerebral lateralization at rest.

Cerebral lateralization is a well-studied topic. However, most of the research to date in functional magnetic resonance imaging (fMRI) has been carried out on hemodynamic fluctuations of voxels, networks, or regions of interest (ROIs). For example, cerebral differences can be revealed by comparing the temporal activation of an ROI in one hemisphere with the corresponding homotopic region in the other hemisphere. While this approach can reveal significant information about cerebral organization, it does not provide information about the full spatiotemporal organization of the hemispheres. The cerebral differences revealed in literature suggest that hemispheres have different spatiotemporal organization in the resting state. In this study, we evaluate cerebral lateralization in the 4D spatiotemporal frequency domain to compare the hemispheres in the context of general activation patterns at different spatial and temporal scales. We use a gender-balanced resting fMRI dataset comprising over 600 healthy subjects ranging in age from 12 to 71, that have previously been studied with a network specific voxel-wise and global analysis of lateralization (Agcaoglu, et al. NeuroImage, 2014). Our analysis elucidates significant differences in the spatiotemporal organization of brain activity between hemispheres, and generally more spatiotemporal fluctuation in the left hemisphere especially in the high spatial frequency bands, and more power in the right hemisphere in the low and middle spatial frequencies. Importantly, the identified effects are not visible in the context of a typical assessment of voxelwise, regional, or even global laterality, thus our study highlights the value of 4D spatiotemporal frequency domain analyses as a complementary and powerful tool for studying brain function.

확산적 흐림의 투과율이 시력과 대비감도에 미치는 영향

Purpose: To investigate the effects of the transmittance of diffusive blurson visual acuity and contrast sensitivity. Methods: Visual acuity and contrast sensitivity were measured by using Optec 6500 in Healthy 30 subjects aged (male 13, female 17) who were recruited from university students. Cataract simulator was used as a tool for diffusive blur. Visual acuity and contrast sensitivity were measured with varying the transmittance of diffusive blur in order to simulate progression of cataract and concentration in fog. Results: Visual acuity was reduced proportionally with decreasing the transmittance of the diffusive blur as follows: . Contrast sensitivity was decreased in all spatial frequencies. Contrast sensitivity in a high spatial frequency band was a greater effect and was off the normal range of contrast sensitivity. The peak of contrast sensitivity was moved in the direction to low frequency. From an intersection point of contrast sensitivity function, we could calculated the transmittance of the diffusive blur being off the normal range and the shift to peak spatial frequency. The peak of contrast sensitivity function was observed to move from 6 to 3 cpd at transmittance of about 78.70%, the contrast sensitivities for all frequencies at transmittance of about 69.71% were deviated from the normal range. Conclusions: The transmittance of diffusive blur causes a reduction in visual acuity and contrast sensitivity, a deviation of normal range of contrast sensitivity, and a shift of peak contrast sensitivity. Therefore the more attention is required when suffering from cataracts or driving in fog.

Capturing developmental shifts in facial identity and expression processing strategies.

Children are widely accepted to process faces differently from adults, with aspects of adult-like processing and neural specialization continuing to develop into early adulthood. Critically, however, little is known about the specific information processing strategies employed as children mature. Here, we investigate the development of face processing strategies in children aged 7-12 years using a reverse correlation approach to determine the visual features driving categorizations of face identity and expression. Specifically, 106 participants (90 children, 16 adults) were trained to identify three unfamiliar faces by name and 200 participants (180 children, 20 adults) were asked to categorize expressions of four basic emotions (fear, sadness, happiness, anger). Across experimental trials we created subsampled versions of these faces by randomly sampling visual information from the images (across different locations and spatial frequency bands) using circularly symmetric apertures ('bubbles'). Participants had to make their categorization decisions based upon only the information directly behind these apertures; the rest was hidden from view. Results revealed that children make less use of the eyes than adults during both tasks. When reading face identity, adults rely on information from the eyes and mouth, particularly in high spatial frequency bands. In contrast, young children make less use of information from the eyes and focus their attention more on the mouth region, with this bias attenuating with age. During expression categorizations we found that relative to adults, children showed a much greater reliance on information from the mouth region when judging fear, sadness and anger, the three emotions that they found most challenging to identify. For happiness, however, which they recognized with near adult-like accuracy, children used information in a more equivalent manner to adults. Across both tasks, these results signal a clear developmental shift in participants' strategic information use when making judgments about key facial characteristics. Meeting abstract presented at VSS 2015.

The impact of an acute social stress on the use of visual information in facial expression categorization

Langner et al. (2009) investigated the impact of social anxiety on the use of visual information in angry vs. neutral emotion recognition using Bubbles (Gosselin & Schyns, 2001). They found that both groups used mostly the same facial features in higher spatial frequencies, but that socially anxious participants relied more than nonanxious participants on low spatial frequencies. Here, we investigated if inducing an acute social stress alters the visual strategies for facial expression categorization using the Bubbles method. Twenty-five men were first submitted to a social stress (i.e. Trier Social Stress Test for Groups) or a control condition (i.e. identical setting save for the socio-evaluative threat component) in a counterbalanced order. Sparse versions of emotional faces (angry, fearful, happy or disgust) were created by sampling facial information at random spatial locations and at four spatial frequency bands. Following the stress induction (or control) phase, participants categorized 500 sparsed stimuli on average. The accuracy was maintained at 62.5% for each facial expression by adjusting the number of bubbles on a trial-by-trial basis. The visual information useful for the task was determined using an analysis procedure that amounts to a multiple linear regression on the bubbles masks and on the participant's response accuracy. This analysis was performed for each condition separately. The facial features useful in high spatial frequencies were similar for both conditions (e.g. the eyes for fear and angry, the mouth for happy). However, the stress condition induced a bias for the use of lower spatial frequencies for angry, fear and happy. For disgust, we replicated Smith et al. (2005) in the control condition, i.e. the importance of the nose winkler. However, in the stress condition, this bias disappeared in favour of the lips. Our results show that an social stress alters the perception of facial expressions of emotions. Meeting abstract presented at VSS 2015. Language: en