Optical Flow Analysis Research Articles

Tissue motion tracking at the edges of a radiation treatment field using local optical flow analysis

This paper investigates the feasibility and accuracy of tracking the motion of an intruding organ-at-risk (OAR) at the edges of a treatment field using a local optical flow analysis of electronic portal images. An intruding OAR was simulated by modifying the portal images obtained by irradiating a programmable phantom's lung tumour. A rectangular treatment aperture was assumed and the edges of the beam's eye view (BEV) were partitioned into clusters/grids according to the width of the multi-leaf collimators (MLC). The optical flow velocities were calculated and the motion accuracy in these clusters was analysed. A velocity error of 0.4 ± 1.4 mm/s with a linearity of 1.04 for tracking an object intruding at 10mm/s (max) was obtained.

Contrast sensitivity and the detection of moving patterns and features

Theories based on optimal sampling by the retina have been widely applied to visual ecology at the level of the optics of the eye, supported by visual behaviour. This leads to speculation about the additional processing that must lie in between-in the brain itself. But fewer studies have adopted a quantitative approach to evaluating the detectability of specific features in these neural pathways. We briefly review this approach with a focus on contrast sensitivity of two parallel pathways for motion processing in insects, one used for analysis of wide-field optic flow, the other for detection of small features. We further use a combination of optical modelling of image blur and physiological recording from both photoreceptors and higher-order small target motion detector neurons sensitive to small targets to show that such neurons operate right at the limits imposed by the optics of the eye and the noise level of single photoreceptors. Despite this, and the limitation of only being able to use information from adjacent receptors to detect target motion, they achieve a contrast sensitivity that rivals that of wide-field motion sensitive pathways in either insects or vertebrates-among the highest in absolute terms seen in any animal.

4D-CT Lung registration using anatomy-based multi-level multi-resolution optical flow analysis and thin-plate splines

The accuracy of 4D-CT registration is limited by inconsistent Hounsfield unit (HU) values in the 4D-CT data from one respiratory phase to another and lower image contrast for lung substructures. This paper presents an optical flow and thin-plate spline (TPS)-based 4D-CT registration method to account for these limitations. The use of unified HU values on multiple anatomy levels (e.g., the lung contour, blood vessels, and parenchyma) accounts for registration errors by inconsistent landmark HU value. While 3D multi-resolution optical flow analysis registers each anatomical level, TPS is employed for propagating the results from one anatomical level to another ultimately leading to the 4D-CT registration. 4D-CT registration was validated using target registration error (TRE), inverse consistency error (ICE) metrics, and a statistical image comparison using Gamma criteria of 1 % intensity difference in 2 mm(3) window range. Validation results showed that the proposed method was able to register CT lung datasets with TRE and ICE values <3 mm. In addition, the average number of voxel that failed the Gamma criteria was <3 %, which supports the clinical applicability of the propose registration mechanism. The proposed 4D-CT registration computes the volumetric lung deformations within clinically viable accuracy.

Primary ciliary dyskinesia assessment by means of optical flow analysis of phase-contrast microscopy images

Primary ciliary dyskinesia implies cilia with defective or total absence of motility, which may result in sinusitis, chronic bronchitis, bronchiectasis and male infertility. Diagnosis can be difficult and is based on an abnormal ciliary beat frequency (CBF) and beat pattern. In this paper, we present a method to determine CBF of isolated cells through the analysis of phase-contrast microscopy images, estimating cilia motion by means of an optical flow algorithm. After having analyzed 28 image sequences (14 with a normal beat pattern and 14 with a dyskinetic pattern), the normal group presented a CBF of 5.2±1.6Hz, while the dyskinetic patients presented a 1.9±0.9Hz CBF. The cutoff value to classify a dyskinetic specimen was set to 3.45Hz (sensitivity 0.86, specificity 0.93). The presented methodology has provided excellent results to objectively diagnose PCD.

Verification of Static Signatures by Optical Flow Analysis

A new approach for static signature verification is presented in this paper. The approach uses optical flow to estimate local stability among signatures. In the enrollment stage, optical flow is used to define a stability model of the genuine signatures for each signer. In the verification stage, the stability between the unknown signature and each one of the reference signatures is estimated and consistency with the stability model of the signer is evaluated. The experimental results, carried out on the signatures in the GPDS database, demonstrate the effectiveness of the new approach.

Action recognition based on statistical analysis from clustered flow vectors

A new technique for action clustering-based human action representation on the basis of optical flow analysis and random sample consensus (RANSAC) method is proposed in this paper. The apparent motion of the human subject with respect to the background is detected and localized by using optical flow analysis. The next task is to characterize the action through the frequent movement of the optical flow points or interest points at different regions of the moving subject. The RANSAC algorithm is used to filter out any unwanted interested points all around the scene and keep only those that are related to that particular subject’s motion. From the remaining salient key interest points, the area of the human body within the frame is estimated. The rectangular area surrounding the human body is then segmented both horizontally and vertically. Now, the percentage of change of interest points in each horizontal and vertical segments from frame to frame is estimated. Similar results are obtained for different persons performing the same action and the corresponding values are averaged for respective segments. The matrix constructed by this strategy is used as a feature vector for that particular action. Similar data are calculated for each block created at the intersections of the horizontal and vertical segments. In addition to these, the change in the position of the person along X- and Y-axes is accumulated for an action and included in the feature vectors. Afterward, for the purpose of recognition using the extracted feature vectors, a distance-based similarity measure and a support vector machine-based classifiers have been exploited. Several combination of the feature vectors is examined. From extensive experimentation upon benchmark motion databases, it is found that the proposed method offers not only a very high degree of accuracy but also computational savings.

El flujo óptico como herramienta para el vídeo-análisis de fenómenos físicos

The optical flow analysis is presented as a useful tool to characterize the movement of a body in an automatic way using a digital video. This technique allows to substitute a complex experimental set-up by a video recording system, which is easily available since the large development in the last years of the multi-media systems. Some results obtained from the analysis of different experiments are shown and the possible use of this kind of technique in a teaching laboratory is discussed.

New measurement method of Poisson's ratio of PVA hydrogels using an optical flow analysis for a digital imaging system

In this paper, we propose a method to observe deformation, to measure strain and to calculate Poisson's ratio of a soft material such as a polyvinyl alcohol hydrogel by means of an optical flow analysis during a tensile test. Compared to the conventional digital image correlation method of which the basic assumption is that the local target region in an image is under linear deformation, an optical flow method can measure local deformation up to the pixel level of the image. In addition, sub-pixel measurements are possible by means of bi-cubic interpolation of each flow vector. To guarantee the accuracy of the optical flow vectors, the scale-invariant feature transform was used. Random sample consensus (RANSAC) for eliminating noisy features and then obtaining a more accurate result was applied. The mean value of Poisson's ratio is 0.4498 and the standard deviation comes to 0.0305. This optical flow method can also be applied to calculate Poisson's ratio of other soft materials.

Optical Flow and Driver’s Kinematics Analysis for State of Alert Sensing

Road accident statistics from different countries show that a significant number of accidents occur due to driver's fatigue and lack of awareness to traffic conditions. In particular, about 60% of the accidents in which long haul truck and bus drivers are involved are attributed to drowsiness and fatigue. It is thus fundamental to improve non-invasive systems for sensing a driver's state of alert. One of the main challenges to correctly resolve the state of alert is measuring the percentage of eyelid closure over time (PERCLOS), despite the driver's head and body movements. In this paper, we propose a technique that involves optical flow and driver's kinematics analysis to improve the robustness of the driver's alert state measurement under pose changes using a single camera with near-infrared illumination. The proposed approach infers and keeps track of the driver's pose in 3D space in order to ensure that eyes can be located correctly, even after periods of partial occlusion, for example, when the driver stares away from the camera. Our experiments show the effectiveness of the approach with a correct eyes detection rate of 99.41%, on average. The results obtained with the proposed approach in an experiment involving fifteen persons under different levels of sleep deprivation also confirm the discriminability of the fatigue levels. In addition to the measurement of fatigue and drowsiness, the pose tracking capability of the proposed approach has potential applications in distraction assessment and alerting of machine operators.

Intraoperative optical flow based tremor evaluation - a feasibility study

Deep brain stimulation as treatment for movement related disorders is a common neurosurgical procedure. Nevertheless the targeting procedure can still be optimized as the clinical outcome resulting from intraoperative test stimulation is in general subjectively evaluated. The aim of the present study was to analyse the feasibility of objective tremor evaluation based on optical flow analysis in videos and to compare the results with in parallel performed acceleration measurements. The results demonstrate the feasibility but as well limitations of the applied setup. Solutions to increase the quality in the future are proposed.

Biologically Inspired Feature Detection Using Cascaded Correlations of off and on Channels

Abstract Flying insects are valuable animal models for elucidating computational processes underlying visual motion detection. For example, optical flow analysis by wide-field motion processing neurons in the insect visual system has been investigated from both behavioral and physiological perspectives [1]. This has resulted in useful computational models with diverse applications [2,3]. In addition, some insects must also extract the movement of their prey or conspecifics from their environment. Such insects have the ability to detect and interact with small moving targets, even amidst a swarm of others [4,5]. We use electrophysiological techniques to record from small target motion detector (STMD) neurons in the insect brain that are likely to subserve these behaviors. Inspired by such recordings, we previously proposed an ‘elementary’ small target motion detector (ESTMD) model that accounts for the spatial and temporal tuning of such neurons and even their ability to discriminate targets against cluttered surrounds [6-8]. However, other properties such as direction selectivity [9] and response facilitation for objects moving on extended trajectories [10] are not accounted for by this model. We therefore propose here two model variants that cascade an ESTMD model with a traditional motion detection model algorithm, the Hassenstein Reichardt ‘elementary motion detector’ (EMD) [11]. We show that these elaborations maintain the principal attributes of ESTMDs (i.e. spatiotemporal tuning and background clutter rejection) while also capturing the direction selectivity observed in some STMD neurons. By encapsulating the properties of biological STMD neurons we aim to develop computational models that can simulate the remarkable capabilities of insects in target discrimination and pursuit for applications in robotics and artificial vision systems.

オプティカルフロー解析によるモーショングラフィックス映像の覚醒度評価手法の検討

オプティカルフロー解析によるモーショングラフィックス映像の覚醒度評価手法の検討

Task contingencies and perceptual strategies shape behavioral effects on neuronal response profiles

We presented optic flow simulating eight directions of self-movement in the ground plane, while monkeys performed delayed match-to-sample tasks, and we recorded dorsal medial superior temporal (MSTd) neuronal activity. Randomly selected sample headings yield smaller test responses to the neuron's preferred heading when it is near the sample's heading direction and larger test responses to the preferred heading when it is far from the sample's heading. Limiting test stimuli to matching or opposite headings suppresses responses to preferred stimuli in both test conditions, whereas focusing on each neuron's preferred vs. antipreferred stimuli enhances responses to the antipreferred stimulus. Match vs. opposite paradigms create bimodal heading profiles shaped by interactions with late delay-period activity. We conclude that task contingencies, determining the prior probabilities of specific stimuli, interact with the monkeys' perceptual strategy for optic flow analysis. These influences shape attentional and working memory effects on the heading direction selectivities and preferences of MSTd neurons.

Investigating lava flow rheology using video analysis and numerical flow models

Lava rheology is a major control on lava flow behavior and a critical parameter in flow simulations, but is very difficult to measure at field conditions or correctly extrapolate from the lab scale. We present a new methodology for investigating lava rheology through a combination of controlled experiments, image analysis and numerical forward modeling. Our experimental setup, part of the Syracuse University Lava Project (http://lavaproject.syr.edu) includes a large furnace capable of melting up to 450kg of basalt, at temperatures well above the basalt liquidus. The lava is poured onto either a tilted bed of sand or a steel channel to produce meter-long flows. This experimental setup is probably the only facility that allows such large scale controlled lava flows made of natural basaltic material.We document the motion of the lava using a high-resolution video camera placed directly above the flows, and the temperature using infrared probes and cameras. After collecting the footage, we analyze the images for lava deformation and compare with numerical forward-models to constrain the rheological parameters and laws which best describe the flowing lava. For the video analysis, we employ the technique of Differential Optical Flow, which uses the time-variations of the spatial gradients of the image intensity to estimate velocity between consecutive frames. An important benefit for using optical flow, compared with other velocimetry methods, is that it outputs a spatially coherent flow field rather than point measurements. We demonstrate that the optical flow results agree with other measures of the flow velocity, and estimate the error due to noise and time-variability to be under 30% of the measured velocity.Our forward-models are calculated by solving the Stokes flow equations on an unstructured finite-element mesh defined using the geometry of the observed flow itself. We explore a range of rheological parameters, including the lava's apparent viscosity, the power-law exponent m and the thermal activation energy. Our measurements of apparent viscosity agree well with predictions of the composition-based Shaw (1972) and GRD model (Giordano, Russell and Dingwell, 2008). We find that for the high-temperature portion of the flow a weakly shear-thinning or Newtonian rheology (m>0.7) with an effective activation energy of B=5500J gives the best fit to the data.Our methodology is the first time that high-resolution optical flow analysis of flowing lava is combined with numerical flow models to constrain rheology. The methodology we present here can be used in field conditions to obtain in-situ information on lava rheology, without physical interaction with the flow and without being limited to point-wise, low strain-rate, local measurements currently available through the use of rotational viscometers in the field.

People Favour Imperfect Catching by Assuming a Stable World

The visual angle that is projected by an object (e.g. a ball) on the retina depends on the object's size and distance. Without further information, however, the visual angle is ambiguous with respect to size and distance, because equal visual angles can be obtained from a big ball at a longer distance and a smaller one at a correspondingly shorter distance. Failure to recover the true 3D structure of the object (e.g. a ball's physical size) causing the ambiguous retinal image can lead to a timing error when catching the ball. Two opposing views are currently prevailing on how people resolve this ambiguity when estimating time to contact. One explanation challenges any inference about what causes the retinal image (i.e. the necessity to recover this 3D structure), and instead favors a direct analysis of optic flow. In contrast, the second view suggests that action timing could be rather based on obtaining an estimate of the 3D structure of the scene. With the latter, systematic errors will be predicted if our inference of the 3D structure fails to reveal the underlying cause of the retinal image. Here we show that hand closure in catching virtual balls is triggered by visual angle, using an assumption of a constant ball size. As a consequence of this assumption, hand closure starts when the ball is at similar distance across trials. From that distance on, the remaining arrival time, therefore, depends on ball's speed. In order to time the catch successfully, closing time was coupled with ball's speed during the motor phase. This strategy led to an increased precision in catching but at the cost of committing systematic errors.

Design and implementation of a distributed fall detection system based on wireless sensor networks

Abstract Pervasive healthcare is one of the most important applications of the Internet of Things (IoT). As part of the IoT, the wireless sensor networks (WSNs) are responsible for sensing the abnormal behavior of the elderly or patients. In this article, we design and implement a fall detection system called SensFall. With the resource restricted sensor nodes, it is vital to find an efficient feature to describe the scene. Based on the optical flow analysis, it can be observed that the thermal energy variation of each sub-region of the monitored region is a salient spatio-temporal feature that characterizes the fall. The main contribution of this study is to develop a feature-specific sensing system to capture this feature so as to detect the occurrence of a fall. In our system, the three-dimensional (3D) object space is segmented into some distinct discrete sampling cells, and pyroelectric infrared (PIR) sensors are employed to detect the variance of the thermal flux within these cells. The hierarchical classifier (two-layer HMMs) is proposed to model the time-varying PIR signal and classify different human activities. We use self-developed PIR sensor nodes mounted on the ceiling and construct a WSN based on ZigBee (802.15.4) protocol. We conduct experiments in a real office environment. The volunteers simulate several kinds of activities including falling, sitting down, standing up from a chair, walking, and jogging. Encouraging experimental results confirm the efficacy of our system.

Cross-range scaling for ISAR via optical flow analysis

For high resolution imaging applications, ISAR can be viewed as a kind of camera which linearly maps the scattering centers from the physical plane to the RD image plane. Thus, the concept and methods of optical flow in computer vision may be introduced for the rotation estimation in ISAR. Furthermore, three kinds of optical flow analysis-based estimators, namely triangle-based, line-based, and image rotation correlation-based estimators, are provided to estimate the coo on the RD image series herein. Because the RD image series is a common output of conventional ISAR and only the scattering center operators are used, the proposed optical flow-based method may also be computationally efficient. Besides, with the increase of scattering centers and images, the high accuracy of cross-range scaling can also be guaranteed. Finally, experimental results are provided to demonstrate the effectiveness of proposed methods based on some measured data of Yak airplane.

CONTOUR INTERPOLATION USING LEVEL-SET ANALYSIS

We propose a novel approach to generate intermediate contours given a sequence of object contours. The proposal unifies shape features through contour curvature analysis and motion between the contours through optic flow analysis. The major contribution of this work is in integrating this shape and image intensity-based contour interpolation scheme in a level-set framework. The interpolated contours between an initial and a target contour act as missing link and establish a path along which contour deformation has taken place. We have shown that for different application domains such as 3D organ visualization (the generation of contours between two spatially apart contours of 2D slice images of a 3D organ), the meteorological applications of tracing, and the path of a developing cyclone (when satellite images are taken at distant time points and the shape of cyclone in between two consecutive satellite images are of interest), the proposal has outperformed the competing approaches.

Bar cells: a novel class of insect feature detector

Bar cells: a novel class of insect feature detector

Characterization of Interictal Epileptiform Discharges with Time-Resolved Cortical Current Maps Using the Helmholtz–Hodge Decomposition

Source estimates performed using a single equivalent current dipole (ECD) model for interictal epileptiform discharges (IEDs) which appear unifocal have proven highly accurate in neocortical epilepsies, falling within millimeters of that demonstrated by electrocorticography. Despite this success, the single ECD solution is limited, best describing sources which are temporally stable. Adapted from the field of optics, optical flow analysis of distributed source models of MEG or EEG data has been proposed as a means to estimate the current motion field of cortical activity, or “cortical flow.” The motion field so defined can be used to identify dynamic features of interest such as patterns of directional flow, current sources, and sinks. The Helmholtz–Hodge Decomposition (HHD) is a technique frequently applied in fluid dynamics to separate a flow pattern into three components: (1) a non-rotational scalar potential U describing sinks and sources, (2) a non-diverging scalar potential A accounting for vortices, and (3) an harmonic vector field H. As IEDs seem likely to represent periods of highly correlated directional flow of cortical currents, the U component of the HHD suggests itself as a way to characterize spikes in terms of current sources and sinks. In a series of patients with refractory epilepsy who were studied with magnetoencephalography as part of their evaluation for possible resective surgery, spike localization with ECD was compared to HHD applied to an optical flow analysis of the same spike. Reasonable anatomic correlation between the two techniques was seen in the majority of patients, suggesting that this method may offer an additional means of characterization of epileptic discharges.