Showing papers on "Motion analysis published in 2015"

Real-Time Fire Detection for Video-Surveillance Applications Using a Combination of Experts Based on Color, Shape, and Motion

Abstract: In this paper, we propose a method that is able to detect fires by analyzing videos acquired by surveillance cameras. Two main novelties have been introduced. First, complementary information, based on color, shape variation, and motion analysis, is combined by a multiexpert system. The main advantage deriving from this approach lies in the fact that the overall performance of the system significantly increases with a relatively small effort made by the designer. Second, a novel descriptor based on a bag-of-words approach has been proposed for representing motion. The proposed method has been tested on a very large dataset of fire videos acquired both in real environments and from the web. The obtained results confirm a consistent reduction in the number of false positives, without paying in terms of accuracy or renouncing the possibility to run the system on embedded platforms.

Hdr-vqm

Abstract: High dynamic range (HDR) signals fundamentally differ from the traditional low dynamic range (LDR) ones in that pixels are related (proportional) to the physical luminance in the scene (i.e. scene-referred). For that reason, the existing LDR video quality measurement methods may not be directly used for assessing quality in HDR videos. To address that, we present an objective HDR video quality measure (HDR-VQM) based on signal pre-processing, transformation, and subsequent frequency based decomposition. Video quality is then computed based on a spatio-temporal analysis that relates to human eye fixation behavior during video viewing. Consequently, the proposed method does not involve expensive computations related to explicit motion analysis in the HDR video signal, and is therefore computationally tractable. We also verified its prediction performance on a comprehensive, in-house subjective HDR video database with 90 sequences, and it was found to be better than some of the existing methods in terms of correlation with subjective scores (for both across sequence and per sequence cases). A software implementation of the proposed scheme is also made publicly available for free download and use. HighlightsThe paper presents one of the first objective method for high dynamic range video quality estimation.It is based on analysis of short term video segments taking into account human viewing behavior.The method described in the paper would be useful in scenarios where HDR video quality needs to be determined in an HDR video chain study.

Toward Automatic Activity Classification and Movement Assessment During a Sports Training Session

Abstract: Motion analysis technologies have been widely used to monitor the potential for injury and enhance athlete performance. However, most of these technologies are expensive, can only be used in laboratory environments, and examine only a few trials of each movement action. In this paper, we present a novel ambulatory motion analysis framework using wearable inertial sensors to accurately assess all of an athlete's activities in real training environment. We first present a system that automatically classifies a large range of training activities using the discrete wavelet transform (DWT) in conjunction with a random forest classifier. The classifier is capable of successfully classifying various activities with up to 98% accuracy. Second, a computationally efficient gradient descent algorithm is used to estimate the relative orientations of the wearable inertial sensors mounted on the shank, thigh, and pelvis of a subject, from which the flexion-extension knee and hip angles are calculated. These angles, along with sacrum impact accelerations, are automatically extracted for each stride during jogging. Finally, normative data are generated and used to determine if a subject's movement technique differed to the normative data in order to identify potential injury-related factors. For the joint angle data, this is achieved using a curve-shift registration technique. It is envisaged that the proposed framework could be utilized for accurate and automatic sports activity classification and reliable movement technique evaluation in various unconstrained environments for both injury management and performance enhancement.

Efficient unsupervised temporal segmentation of human motion

Abstract: This work introduces an efficient method for fully automatic temporal segmentation of human motion sequences and similar time series. The method relies on a neighborhood graph to partition a given data sequence into distinct activities and motion primitives according to self-similar structures given in that input sequence. In particular, the fast detection of repetitions within the discovered activity segments is a crucial problem of any motion processing pipeline directed at motion analysis and synthesis. The same similarity information in the neighborhood graph is further exploited to cluster these primitives into larger entities of semantic significance. The elements subject to this classification are then used as prior for estimating the same target values for entirely unknown streams of data.The technique makes no assumptions about the motion sequences at hand and no user interaction is required for the segmentation or clustering. Tests of our techniques are conducted on the CMU and HDM05 motion capture databases demonstrating the capability of our system handling motion segmentation, clustering, motion synthesis and transfer-of-label problems in practice - the latter being an optional step which relies on the preexistence of a small set of labeled data.

Moving object tracking using optical flow and motion vector estimation

Abstract: Moving object detection and tracking is an evolving research field due to its wide applications in traffic surveillance, 3D reconstruction, motion analysis (human and non-human), activity recognition, medical imaging etc. However real time object tracking is a challenging task due to dynamic tacking environment and different limiting parameters like view point, anthropometric variation, dimensions of an object, cluttered background, camera motions, occlusion etc. In this paper, we have developed new object detection and tracking algorithm which makes use of optical flow in conjunction with motion vector estimation for object detection and tracking in a sequence of frames. The optical flow gives valuable information about the object movement even if no quantitative parameters are computed. The motion vector estimation technique can provide an estimation of object position from consecutive frames which increases the accuracy of this algorithm and helps to provide robust result irrespective of image blur and cluttered background. The use of median filter with this algorithm makes it more robust in the presence of noise. The developed algorithm is applied to wide range of standard and real time datasets with different illumination (indoor and outdoor), object speed etc. The obtained results indicates that the developed algorithm over performs over conventional methods and state of art methods of object tracking.

Integrated sensor and video motion analysis method

Abstract: A method that integrates sensor data and video analysis to analyze object motion. Motion capture elements generate motion sensor data for objects of interest, and cameras generate video of these objects. Sensor data and video data are synchronized in time and aligned in space on a common coordinate system. Sensor fusion is used to generate motion metrics from the combined and integrated sensor data and video data. Integration of sensor data and video data supports robust detection of events, generation of video highlight reels or epic fail reels augmented with metrics that show interesting activity, and calculation of metrics that exceed the individual capabilities of either sensors or video analysis alone.

Event analysis system

Abstract: Enables event analysis from sensors including environmental, physiological and motion capture sensors. Also enables displaying information based on events recognized using sensor data associated with a user, piece of equipment or based on previous motion analysis data from the user or other user(s) or other sensors. Enables intelligent analysis, synchronization, and transfer of generally concise event videos synchronized with motion data from motion capture sensor(s) coupled with a user or piece of equipment. Enables creating, transferring, obtaining, and storing concise event videos generally without non-event video. Events stored in the database identifies trends, correlations, models, and patterns in event data. Greatly saves storage and increases upload speed by uploading event videos and avoiding upload of non-pertinent portions of large videos. Creates highlight and fail reels filtered by metrics and can sort by metric. Compares motion trajectories of users and objects to optimally efficient trajectories, and to desired trajectories.

Evaluation of a portable markerless finger position capture device: accuracy of the Leap Motion controller in healthy adults.

Abstract: Although motion analysis is frequently employed in upper limb motor assessment (e.g. visually-guided reaching), they are resource-intensive and limited to laboratory settings. This study evaluated the reliability and accuracy of a new markerless motion capture device, the Leap Motion controller, to measure finger position. Testing conditions that influence reliability and agreement between the Leap and a research-grade motion capture system were examined. Nine healthy young adults pointed to 15 targets on a computer screen under two conditions: (1) touching the target (touch) and (2) 4 cm away from the target (no-touch). Leap data was compared to an Optotrak marker attached to the index finger. Across all trials, root mean square (RMS) error of the Leap system was 17.30 ± 9.56 mm (mean ± SD), sampled at 65.47 ± 21.53 Hz. The % viable trials and mean sampling rate were significantly lower in the touch condition (44% versus 64%, p < 0.001; 52.02 ± 2.93 versus 73.98 ± 4.48 Hz, p = 0.003). While linear correlations were high (horizontal: r2 = 0.995, vertical r2 = 0.945), the limits of agreement were large (horizontal: −22.02 to +26.80 mm, vertical: −29.41 to +30.14 mm). While not as precise as more sophisticated optical motion capture systems, the Leap Motion controller is sufficiently reliable for measuring motor performance in pointing tasks that do not require high positional accuracy (e.g. reaction time, Fitt’s, trails, bimanual coordination).

Abnormal gait detection with RGB-D devices using joint motion history features

Abstract: Human gait has become of special interest to health professionals and researchers in recent years, not only due to its relation to a person's quality of life and personal autonomy, but also due to the involved cognitive process, since deviation from normal gait patterns can also be associated to neurological diseases. Vision-based abnormal gait detection can provide support to current human gait analysis procedures providing quantitative and objective metrics that can assist the evaluation of the geriatrician, while at the same time providing technical advantages, such as low intrusiveness and simplified setups. Furthermore, recent advances in RGB-D devices allow to provide low-cost solutions for 3D human body motion analysis. In this sense, this work presents a method for abnormal gait detection relying on skeletal pose representation based on depth data. A novel spatio-temporal feature is presented that provides a representation of a set of consecutive skeletons based on the 3D location of the skeletal joints and the motion's age. The corresponding feature sequences are learned using a machine learning method, namely BagOfKeyPoses. Experimentation with different datasets and evaluation methods shows that reliable detection of abnormal gait is obtained and, at the same time, an outstandingly high temporal performance is provided.

Reliability and accuracy of a goniometer mobile device application for video measurement of the functional movement screen deep squat test.

Abstract: Background/Purpose The squat is a fundamental movement of many athletic and daily activities. Methods to clinically assess the squat maneuver range from simple observation to the use of sophisticated equipment. The purpose of this study was to examine the reliability of Coach's Eye (TechSmith Corp), a 2‐dimensional (2D) motion analysis mobile device application (app), for assessing maximal sagittal plane hip, knee, and ankle motion during a functional movement screen deep squat, and to compare range of motion values generated by it to those from a Vicon (Vicon Motion Systems Ltd) 3‐dimensional (3D) motion analysis system.

A Direct PCA-Based Approach for Real-Time Description of Physiological Organ Deformations

Abstract: Dynamic magnetic resonance (MR)-imaging can provide functional and positional information in real-time, which can be conveniently used online to control a cancer therapy, e.g., using high intensity focused ultrasound or radio therapy. However, a precise real-time correction for motion is fundamental in abdominal organs to ensure an optimal treatment dose associated with a limited toxicity in nearby organs at risk. This paper proposes a real-time direct principal component analysis (PCA)-based technique which offers a robust approach for motion estimation of abdominal organs and allows correcting motion related artifacts. The PCA was used to detect spatio-temporal coherences of the periodic organ motion in a learning step. During the interventional procedure, physiological contributions were characterized quantitatively using a small set of parameters. A coarse-to-fine resolution scheme is proposed to improve the stability of the algorithm and afford a predictable constant latency of 80 ms. The technique was evaluated on 12 free-breathing volunteers and provided an improved real-time description of motion related to both breathing and cardiac cycles. A reduced learning step of 10 s was sufficient without any need for patient-specific control parameters, rendering the method suitable for clinical use.

Simultaneous Vision-Based Shape and Motion Analysis of Cells Fast-Flowing in a Microchannel

Abstract: This paper proposes a novel concept for simultaneous cell shape and motion analysis in fast microchannel flows by implementing a multiobject feature extraction algorithm on a frame-straddling high-speed vision platform. The system can synchronize two camera inputs with the same view with only a tiny time delay on the sub-microsecond timescale. Real-time video processing is performed in hardware logic by extracting the moment features of multiple cells in 512 $\,\times\,$ 256 images at 4000 fps for the two camera inputs and their frame-straddling time can be adjusted from 0 to 0.25 ms in 9.9 ns steps. By setting the frame-straddling time in a certain range to avoid large image displacements between the two camera inputs, our frame-straddling high-speed vision platform can perform simultaneous shape and motion analysis of cells in fast microchannel flows of 1 m/s or greater. The results of real-time experiments conducted to analyze the deformabilities and velocities of sea urchin egg cells fast-flowing in microchannels verify the efficacy of our vision-based cell analysis system.

Magnetic resonance imaging-guided versus surrogate-based motion tracking in liver radiation therapy: a prospective comparative study.

Abstract: Purpose This study applied automatic feature detection on cine–magnetic resonance imaging (MRI) liver images in order to provide a prospective comparison between MRI-guided and surrogate-based tracking methods for motion-compensated liver radiation therapy. Methods and Materials In a population of 30 subjects (5 volunteers plus 25 patients), 2 oblique sagittal slices were acquired across the liver at high temporal resolution. An algorithm based on scale invariant feature transform (SIFT) was used to extract and track multiple features throughout the image sequence. The position of abdominal markers was also measured directly from the image series, and the internal motion of each feature was quantified through multiparametric analysis. Surrogate-based tumor tracking with a state-of-the-art external/internal correlation model was simulated. The geometrical tracking error was measured, and its correlation with external motion parameters was also investigated. Finally, the potential gain in tracking accuracy relying on MRI guidance was quantified as a function of the maximum allowed tracking error. Results An average of 45 features was extracted for each subject across the whole liver. The multi-parametric motion analysis reported relevant inter- and intrasubject variability, highlighting the value of patient-specific and spatially-distributed measurements. Surrogate-based tracking errors (relative to the motion amplitude) were were in the range 7% to 23% (1.02-3.57mm) and were significantly influenced by external motion parameters. The gain of MRI guidance compared to surrogate-based motion tracking was larger than 30% in 50% of the subjects when considering a 1.5-mm tracking error tolerance. Conclusions Automatic feature detection applied to cine-MRI allows detailed liver motion description to be obtained. Such information was used to quantify the performance of surrogate-based tracking methods and to provide a prospective comparison with respect to MRI-guided radiation therapy, which could support the definition of patient-specific optimal treatment strategies.

Improved Leg Tracking Considering Gait Phase and Spline-Based Interpolation during Turning Motion in Walk Tests.

Abstract: Falling is a common problem in the growing elderly population, and fall-risk assessment systems are needed for community-based fall prevention programs. In particular, the timed up and go test (TUG) is the clinical test most often used to evaluate elderly individual ambulatory ability in many clinical institutions or local communities. This study presents an improved leg tracking method using a laser range sensor (LRS) for a gait measurement system to evaluate the motor function in walk tests, such as the TUG. The system tracks both legs and measures the trajectory of both legs. However, both legs might be close to each other, and one leg might be hidden from the sensor. This is especially the case during the turning motion in the TUG, where the time that a leg is hidden from the LRS is longer than that during straight walking and the moving direction rapidly changes. These situations are likely to lead to false tracking and deteriorate the measurement accuracy of the leg positions. To solve these problems, a novel data association considering gait phase and a Catmull–Rom spline-based interpolation during the occlusion are proposed. From the experimental results with young people, we confirm that the proposed methods can reduce the chances of false tracking. In addition, we verify the measurement accuracy of the leg trajectory compared to a three-dimensional motion analysis system (VICON).

BioKin: an ambulatory platform for gait kinematic and feature assessment.

Abstract: A platform to move gait analysis, which is normally restricted to a clinical environment in a well-equipped gait laboratory, into an ambulatory system, potentially in non-clinical settings is introduced. This novel system can provide functional measurements to guide therapeutic interventions for people requiring rehabilitation with limited access to such gait laboratories. BioKin system consists of three layers: a low-cost wearable wireless motion capture sensor, data collection and storage engine, and the motion analysis and visualisation platform. Moreover, a novel limb orientation estimation algorithm is implemented in the motion analysis platform. The performance of the orientation estimation algorithm is validated against the orientation results from a commercial optical motion analysis system and an instrumented treadmill. The study results demonstrate a root-mean-square error less than 4° and a correlation coefficient more than 0.95 when compared with the industry standard system. These results indicate that the proposed motion analysis platform is a potential addition to existing gait laboratories in order to facilitate gait analysis in remote locations.

Unsupervised detection of nonlinearity in motion using weighted average of non-extensive entropies

Abstract: Motion estimation and motion analysis have an important role to play for detecting abnormal motion in surveillance videos. In this paper, we propose to use the non-extensive entropy to detect any unnaturalness in the motion over correlated video frames since it has already been proved to represent the correlated textures successfully. To achieve this end, we utilize the temporal correlation property of motion vectors over three consecutive frames to detect any motion disturbance using a weighted average of the non-extensive entropies. It is proved by the experimental results on the state-of-the-art database that the non-extensive entropy is most apt for detecting any disturbance in the continuance of motion vectors in between frames. The advantage of our approach is that no training period or normalcy reference is required since a relative disturbance in the magnitudes of motion vectors over a three-frame window gives an alarm.

On-the-fly hand detection training with application in egocentric action recognition

Abstract: We propose a novel approach to segment hand regions in egocentric video that requires no manual labeling of training samples. The user wearing a head-mounted camera is prompted to perform a simple gesture during an initial calibration step. A combination of color and motion analysis that exploits knowledge of the expected gesture is applied on the calibration video frames to automatically label hand pixels in an unsupervised fashion. The hand pixels identified in this manner are used to train a statistical-model-based hand detector. Superpixel region growing is used to perform segmentation refinement and improve robustness to noise. Experiments show that our hand detection technique based on the proposed on-the-fly training approach significantly outperforms state-of-the-art techniques with respect to accuracy and robustness on a variety of challenging videos. This is due primarily to the fact that training samples are personalized to a specific user and environmental conditions. We also demonstrate the utility of our hand detection technique to inform an adaptive video sampling strategy that improves both computational speed and accuracy of egocentric action recognition algorithms. Finally, we offer an egocentric video dataset of an insulin self-injection procedure with action labels and hand masks that can serve towards future research on both hand detection and egocentric action recognition.

A Simple Method to Calibrate Kinematical Invariants: Application to Overhead Throwing

Abstract: The aim of this paper is to present a simple calibration method aimed at optimizing the kinematical invariants of a whole body motion capture model, meaning limb lengths and some of the marker placements. A case study and preliminary results are presented and give encouraging insights about the generalized use of such a method in motion analysis in sports.

A Local Feature based on Lagrangian Measures for Violent Video Classification

Abstract: Lagrangian theory provides a diverse set of tools for continuous motion analysis. Existing work shows the applicability of Lagrangian methods for video analysis in several aspects. In this paper we want to utilize the concept of Lagrangian measures to detect violent scenes. Therefore we propose a local feature based on the SIFT algorithm that incooperates appearance and Lagrangian based motion models. We will show that the temporal interval of the used motion information is a crucial aspect and study its influence on the classification performance. The proposed LaSIFT feature outperforms other state-of-the-art local features, in particular in uncontrolled realistic video data. We evaluate our algorithm with a bag-of-word approach. The experimental results show a significant improvement over the state-of-the-art on current violence detection datasets, i.e. Crowd Violence, Hockey Fight.

Evaluation of functional methods for human movement modelling.

Abstract: PURPOSE: accurate assessment of human joint parameters is a critical issue for the quantitative movement analysis, due to a direct influence on motion patterns. In this study three different known functional methods are experimentally compared to identify knee joint kinematics for further gait and motion analysis purposes. METHODS: taking into account the human knee physiology complexity, within its roto-translation, the study is conducted on a lower limb mechanical analogue with a polycentric hinge-based kinematic model. The device mimics a joint with a mobile axis of rotation whose position is definable. Sets of reflective markers are placed on the dummy and flexion-extension movements are imposed to the shank segment. Marker positions are acquired using an optoelectronic motion capture system (Vicon 512). RESULTS: acquired markers' positions are used as input data to the three functional methods considered. These ones approximate the polycentric knee joint with a fixed single axis model. Different ranges of motion and number of markers are considered for each functional method. RESULTS are presented through the evaluation of accuracy and precision concerning both misalignment and distance errors between the estimated axis of rotation and the instantaneous polycentric one, used as reference. CONCLUSION: the study shows the feasibility of the identification of joint parameters with functional approaches applied on a polycentric mechanism, differently from those usually conceived by the reviewed algorithms. Moreover, it quantifies and compares the approximation errors using different algorithms, by varying number and position of markers, as well ranges of motion. Language: en

Temporal segmentation and keyframe selection methods for user-generated video search-based annotation

Abstract: Video temporal segmentation and keyframe selection approaches for User Generated Video (UGV)Hierarchical Hidden Markov Models applied to camera motion analysis to detect motion patterns and temporally segment the videoEvaluation of the influence of camera motion over the performance of automatic UGV annotation systemsTwo datasets for User Generated Video have been developed and made publicly available In this paper we propose a temporal segmentation and a keyframe selection method for User-Generated Video (UGV) Since UGV is rarely structured in shots and usually user's interest are revealed through camera movements, a UGV temporal segmentation system has been proposed that generates a video partition based on a camera motion classification Motion-related mid-level features have been suggested to feed a Hierarchical Hidden Markov Model (HHMM) that produces a user-meaningful UGV temporal segmentation Moreover, a keyframe selection method has been proposed that picks a keyframe for fixed-content camera motion patterns such as zoom, still, or shake and a set of keyframes for varying-content translation patternsThe proposed video segmentation approach has been compared to a state-of-the-art algorithm, achieving 8% performance improvement in a segmentation-based evaluation Furthermore, a complete search-based UGV annotation system has been developed to assess the influence of the proposed algorithms on an end-user task To that purpose, two UGV datasets have been developed and made available online Specifically, the relevance of the considered camera motion types has been analyzed for these two datasets, and some guidelines are given to achieve the desired performance-complexity tradeoff The keyframe selection algorithm for varying-content translation patterns has also been assessed, revealing a notable contribution to the performance of the global UGV annotation system Finally, it has been shown that the UGV segmentation algorithm also produces improved annotation results with respect to a fixed-rate keyframe selection baseline or a traditional method relying on frame-level visual features

Feasibility of Using Low-Cost Motion Capture for Automated Screening of Shoulder Motion Limitation after Breast Cancer Surgery.

Abstract: Objective To determine if a low-cost, automated motion analysis system using Microsoft Kinect could accurately measure shoulder motion and detect motion impairments in women following breast cancer surgery. Design Descriptive study of motion measured via 2 methods. Setting Academic cancer center oncology clinic. Participants 20 women (mean age = 60 yrs) were assessed for active and passive shoulder motions during a routine post-operative clinic visit (mean = 18 days after surgery) following mastectomy (n = 4) or lumpectomy (n = 16) for breast cancer. Interventions Participants performed 3 repetitions of active and passive shoulder motions on the side of the breast surgery. Arm motion was recorded using motion capture by Kinect for Windows sensor and on video. Goniometric values were determined from video recordings, while motion capture data were transformed to joint angles using 2 methods (body angle and projection angle). Main Outcome Measure Correlation of motion capture with goniometry and detection of motion limitation. Results Active shoulder motion measured with low-cost motion capture agreed well with goniometry (r = 0.70–0.80), while passive shoulder motion measurements did not correlate well. Using motion capture, it was possible to reliably identify participants whose range of shoulder motion was reduced by 40% or more. Conclusions Low-cost, automated motion analysis may be acceptable to screen for moderate to severe motion impairments in active shoulder motion. Automatic detection of motion limitation may allow quick screening to be performed in an oncologist's office and trigger timely referrals for rehabilitation.

Correlation-optimized time warping for motion

Abstract: Retrieval and comparative editing/modeling of motion data require temporal alignment. In other words, for such processes to perform accurately, critical features of motion sequences need to occur simultaneously. In this paper, we propose correlation-optimized time warping (CoTW) for aligning motion data. CoTW utilizes a correlation-based objective function for characterizing alignment. The method solves an optimization problem to determine the optimum warping degree for different segments of the sequence. Using segment-wise interpolated warping, smooth motion trajectories are achieved that can be readily used for animation. Our method allows for manual tuning of the parameters, resulting in high customizability with respect to the number of actions in a single sequence as well as spatial regions of interest within the character model. Moreover, measures are taken to reduce distortion caused by over-warping. The framework also allows for automatic selection of an optimum reference when multiple sequences are available. Experimental results demonstrate the very accurate performance of CoTW compared to other techniques such as dynamic time warping, derivative dynamic time warping and canonical time warping. The mentioned customization capabilities are also illustrated.