Gait Representation Research Articles

Change Energy Image for Gait Recognition: An Approach Based on Symbolic Representation

Gait can be identified by observing static and dynamic parts of human body. In this paper a variant of gait energy image called change energy images (CEI) are generated to capture detailed static and dynamic information of human gait. Radon transform is applied to CEI in four different directions (vertical, horizontal and two opposite cross sections) considering four different angles to compute discriminative feature values. The extracted features are represented in the form of interval -valued type symbolic data. The proposed method is capable of recognizing an individual when he/she have variations in their gait due to different clothes they wear, in different normal conditions and carrying a bag. A similarity measure suitable for the proposed gait representation is explored for the purpose of establishing similarity match for gait recognition. Experiments are conducted on CASIA database B and the results have shown better recognition performance compared to some of the existing methods.

The Influence of Visual Feedback on the Mental Representation of Gait in Patients with THR: A New Approach for an Experimental Rehabilitation Strategy

Due to total hip replacement (THR), patients reveal abnormal gait patterns which post-operative do often not return to "normal". The restoration towards normal gait reduces stress on the adjacent joints which consequently reduces risk of osteoarthrosis development. Motor-performance is related to the structure of the movement in long-term memory, thus it seems to be essential to imprint correct gait patterns in there. Mental representation structures can develop over the course of training and visual feedback presumably helps regaining a better representation of gait in long-term memory. The purpose of this study is to evaluate the effect of visual feedback on mental representation in patients with THR. In a randomized controlled trial, 20 women (57 ± 6 years) with THR have been enrolled. Subjects were randomly assigned to a control group (CG) or intervention group (IG). Additionally to inpatient treatment, all subjects participated in a standardized gait training including either an intervention based on verbal information from a physiotherapist (CG) or an intervention based on real-time visual feedback (IG). Mental representation was measured in pre-test and post-test using the structure-dimensional analysis. Results indicate significant improvements in mental representation of gait in the post-test only in IG, suggesting that beneficial effects were provoked by visual feedback.

Real Time Video Surveillance System

The significance of Real Time Video Surveillance System has been elevated due to increased importance in safety and security. The Real Time Video Surveillance System is capable of detecting, classifying and tracking objects of interest (human), by recognizing motion of a walking person in video, which can lead to a useful system. This is done by generating notification to an authoritative person by displaying the recorded face of the threat. Gait Analysis helps us to identify a person by their walking style, which also helps for individual recognition i.e. for performing feature extraction. To accomplish this we have used a new spatio-temporal gait representation called as Gait Energy Image (GEI). GEI represents a human motion sequence in a single image while preserving temporal information. For human recognition, statistical GEI feature matching is used to overcome limitation of training templates. Also to reduce dimensionality problem of GEI’s, we have used two approaches, namely Principal Component Analysis (PCA) and its variants Multiple Discriminant Analysis (MDA). These approaches helped us to achieve good performance over the existing approaches.

Use of Gait Energy Image in Implementation of Real Time Video Surveillance System

The significance of Real Time video surveillance System has been raised up due to increased importance in safety and security. The Proposed Real Time video surveillance system is capable of detecting, classifying and tracking objects of interest (human). This can be done by using Gait Analysis for tracking scenarios and generating notification to authoritative person. Gait Analysis helps us to identify people by their walking style. Because of this exquisite feature we used gait recognition in proposed real time video surveillance system. In this proposed system for depicting human walking properties for individual recognition i.e. for performing feature extraction we have used a new spatio-temporal gait representation called as Gait Energy Image (GEI). GEI represents a human motion sequence in a single image while preserving temporal information. For human recognition we have used Statistical GEI feature matching, wherein to reduce dimensionality problem of GEI's, we used two approaches they are Principal Component Analysis (PCA) and its variants Multiple Discriminant Analysis (MDA).

Gait recognition based on Gabor wavelets and modified gait energy image for human identification

This paper proposes a method for recognizing human identity using gait features based on Gabor wavelets and modified gait energy images (GEIs). Identity recognition by gait generally involves gait representation, extraction, and classification. In this work, a modified GEI convolved with an ensemble of Gabor wavelets is proposed as a gait feature. Principal component analysis is then used to project the Gabor-wavelet-based gait features into a lower-dimension feature space for subsequent classification. Finally, support vector machine classifiers based on a radial basis function kernel are trained and utilized to recognize human identity. The major contributions of this paper are as follows: (1) the consideration of the shadow effect to yield a more complete segmentation of gait silhouettes; (2) the utilization of motion estimation to track people when walkers overlap; and (3) the derivation of modified GEIs to extract more useful gait information. Extensive performance evaluation shows a great improvement of recognition accuracy due to the use of shadow removal, motion estimation, and gait representation using the modified GEIs and Gabor wavelets.

Effective part-based gait identification using frequency-domain gait entropy features

Gait identification task becomes more difficult due to the change of appearance by different cofactors (e.g., shoe, surface, carrying, view, and clothing). The cofactors may affect some parts of gait while other parts remain unchanged and can be used for recognition. We propose a robust technique to define which parts are more effective and which parts are less effective for cofactors like clothing, carrying objects etc. To find out the effective body parts, the whole body is divided into small segments where each segment is a single row in this paper. Based on positive and negative effect of each segment, three most effective parts and two less effective parts are defined. Usually, the dynamic areas (e.g., legs, arms swing) are comparatively less affected than static areas (e.g., torso) for different cofactors in appearance based gait representation. To give more emphasis on dynamic areas and less on static areas, frequency-domain gait entropy termed as EnDFT representation is computed and used as gait features. Experiments are conducted on two comprehensive benchmarking databases: The OU-ISIR Gait Database, the Treadmill dataset B with clothing variations and CASIA Gait Database, Dataset B with clothing and carrying conditions. The proposed method shows better results in comparison with other existing gait recognition approaches.

Gait-based Gender Classification Considering Resampling and Feature Selection

 Abstract—Two intrinsic data characteristics that arise in many domains are the class imbalance and the high dimensionality, which pose new challenges that should be addressed. When using gait for gender classification, benchmarking public databases and renowned gait representations lead to these two problems, but they have not been jointly studied in depth. This paper is a preliminary study that pursues to investigate the benefits of using several techniques to tackle the aforementioned problems either singly or in combination, and also to evaluate the order of application that leads to the best classification performance. Experimental results show the importance of jointly managing both problems for gaitbased gender classification. In particular, it seems that the best strategy consists of applying resampling followed by feature selection. 

Two-layer dual gait generative models for human motion estimation from a single camera

This paper presents a two-layer gait representation framework for video-based human motion estimation that extends our recent dual gait generative models, visual gait generative model (VGGM) and kinematic gait generative model (KGGM), with a new capability of part-whole gait modeling. Specifically, the idea of gait manifold learning is revisited to capture the gait variability among different individuals at both whole and part levels. A key issue is the selection of an appropriate distance metric to evaluate the dissimilarity between two gaits (either at whole or part levels) that determines an optimal manifold topology. Several metrics are studied and compared in terms of their effectiveness for gait manifold learning at both whole and part levels. This work involves one whole-based and two part-level gait manifolds by which three pairs of KGGM and VGGM can be learned and integrated for part-whole gait modeling. Moreover, a two-stage Monte Carlo Markov Chain (MCMC) inference algorithm is developed for video-based part-whole motion estimation. The proposed algorithm is tested on the HumanEva data and reaches state-of-art results.

A probabilistic image-weighting scheme for robust silhouette-based gait recognition

Many gait recognition methods use silhouettes as a feature due to their simplicity and effectiveness. However, silhouette-based gait recognition algorithms have the drawback of performance degradation when the silhouette images are corrupted. To solve this problem, this paper proposes a new gait representation method by emphasizing the noise-free silhouettes while suppressing the corrupted ones. The probabilistic support vector machine (PSVM) is employed to weigh the silhouette images according to quality and to construct a new gait representation for robust recognition. Experiments are conducted with the CASIA and SOTON databases, and the proposed method makes silhouette-based gait recognition as reliable biometrics.

The Effect of Time on Gait Recognition Performance

Many studies have shown that it is possible to recognize people by the way they walk. However, there are a number of covariate factors that affect recognition performance. The time between capturing the gallery and the probe has been reported to affect recognition the most. To date, no study has isolated the effect of time, irrespective of other covariates. Here, we present the first principled study that examines the effect of elapsed time on gait recognition. Using empirical evidence we show for the first time that elapsed time does not affect recognition significantly in the short-medium term. This finding challenges the existing view in the literature that time significantly affects gait recognition. We employ existing gait representations on a novel dataset captured specifically for this study. By controlling the clothing worn by the subjects and the environment, a Correct Classification Rate (CCR) of 95% has been achieved over the longest time period yet considered for gait on the largest ever temporal dataset. Our results show that gait can be used as a reliable biometric over time and at a distance if we were able to control all other factors such as clothing, footwear etc. We have also investigated the effect of different type of clothes, variations in speed and footwear on the recognition performance. The purpose of these experiments is to provide an indication of why previous studies (employing the same techniques as this study) have achieved significantly lower recognition performance over time. Our experimental results show that clothing and other covariates have been confused with elapsed time previously in the literature. We have demonstrated that clothing drastically affects the recognition performance regardless of elapsed time and significantly more than any of the other covariates that we have considered here.

SSM-HPC: Front View Gait Recognition Using Spherical Space Model with Human Point Clouds

In this paper, we propose a novel gait recognition framework - Spherical Space Model with Human Point Clouds (SSM-HPC) to recognize front view of human gait. A new gait representation - Marching in Place (MIP) gait is also introduced which preserves the spatiotemporal characteristics of individual gait manner. In comparison with the previous studies on gait recognition which usually use human silhouette images from image sequences, this research applies three dimensional (3D) point clouds data of human body obtained from stereo camera. The proposed framework exhibits gait recognition rates superior to those of other gait recognition methods.

Gait Recognition With Shifted Energy Image and Structural Feature Extraction

In this paper, we present a novel and efficient gait recognition system. The proposed system uses two novel gait representations, i.e., the shifted energy image and the gait structural profile, which have increased robustness to some classes of structural variations. Furthermore, we introduce a novel method for the simulation of walking conditions and the generation of artificial subjects that are used for the application of linear discriminant analysis. In the decision stage, the two representations are fused. Thorough experimental evaluation, conducted using one traditional and two new databases, demonstrates the advantages of the proposed system in comparison with current state-of-the-art systems.

Gait recognition using Pose Kinematics and Pose Energy Image

Many of the existing gait recognition approaches represent a gait cycle using a single 2D image called Gait Energy Image (GEI) or its variants. Since these methods suffer from lack of dynamic information, we model a gait cycle using a chain of key poses and extract a novel feature called Pose Energy Image (PEI). PEI is the average image of all the silhouettes in a key pose state of a gait cycle. By increasing the resolution of gait representation, more detailed dynamic information can be captured. However, processing speed and space requirement are higher for PEI than the conventional GEI methods. To overcome this shortcoming, another novel feature named as Pose Kinematics is introduced, which represents the percentage of time spent in each key pose state over a gait cycle. Although the Pose Kinematics based method is fast, its accuracy is not very high. A hierarchical method for combining these two features is, therefore, proposed. At first, Pose Kinematics is applied to select a set of most probable classes. Then, PEI is used on these selected classes to get the final classification. Experimental results on CMU's Mobo and USF's HumanID data set show that the proposed approach outperforms existing approaches.

P6.15 The basiogram: a plot representation of gait

P6.15 The basiogram: a plot representation of gait

Gait flow image: A silhouette-based gait representation for human identification

In this paper, we propose a novel gait representation-gait flow image (GFI) for use in gait recognition. This representation will further improve recognition rates. The basis of GFI is the binary silhouette sequence. GFI is generated by using an optical flow field without constructing any model. The performance of the proposed representation was evaluated and compared with the other representations, such as gait energy image (GEI), experimentally on the USF data set. The USF data set is a public data set in which the image sequences were captured outdoors. The experimental results show that the proposed representation is efficient for human identification. The average recognition rate of GFI is better than that of the other representations in direct matching and dimensional reduction approaches. In the direct matching approach, GFI achieved an average identification rate 42.83%, which is better than GEI by 3.75%. In the dimensional reduction approach, GFI achieved an average identification rate 43.08%, which is better than GEI by 1.5%. The experimental result showed that GFI is stronger in resisting the difference of the carrying condition compared with other gait representations.

Special Issue on New Advances in Video-Based Gait Analysis and Applications: Challenges and Solutions

The six articles in this special issue span a variety of topics in terms of gait representation and analysis for different applications.

Dual Gait Generative Models for Human Motion Estimation From a Single Camera

This paper presents a general gait representation framework for video-based human motion estimation. Specifically, we want to estimate the kinematics of an unknown gait from image sequences taken by a single camera. This approach involves two generative models, called the kinematic gait generative model (KGGM) and the visual gait generative model (VGGM), which represent the kinematics and appearances of a gait by a few latent variables, respectively. The concept of gait manifold is proposed to capture the gait variability among different individuals by which KGGM and VGGM can be integrated together, so that a new gait with unknown kinematics can be inferred from gait appearances via KGGM and VGGM. Moreover, a new particle-filtering algorithm is proposed for dynamic gait estimation, which is embedded with a segmental jump-diffusion Markov Chain Monte Carlo scheme to accommodate the gait variability in a long observed sequence. The proposed algorithm is trained from the Carnegie Mellon University (CMU) Mocap data and tested on the Brown University HumanEva data with promising results.

Recursive spatiotemporal subspace learning for gait recognition

In this paper, we propose a new gait recognition method using recursive spatiotemporal subspace learning. In the first stage, periodic dynamic feature of gait over time is extracted by Principal Component Analysis (PCA) and gait sequences are represented in the form of Periodicity Feature Vector (PFV). In the second stage, shape feature of gait over space is extracted by Discriminative Locality Alignment (DLA) based on the PFV representation of gait sequences. After the recursive subspace learning, gait sequence data is compressed into a very compact vector named Gait Feature Vector (GFV) which is used for individual recognition. Compared to other gait recognition methods, GFV is an effective representation of gait because the recursive spatiotemporal subspace learning technique extracts both the shape features and the dynamic features. And at the same time, representing gait sequences in PFV form is an efficient way to save storage space and computational time. Experimental result shows that the proposed method achieves highly competitive performance with respect to the published gait recognition approaches on the USF HumanID gait database.

Active energy image plus 2DLPP for gait recognition

This paper proposes a novel active energy image (AEI) method for gait recognition. Existing human gait feature representation methods, however, usually suffer from low quality of human silhouettes and insufficient dynamic characteristics. To this end, we apply the proposed AEI for gait representation. Given a gait silhouette sequence, we first extract the active regions by calculating the difference of two adjacent silhouette images, and construct an AEI by accumulating these active regions. Then, we project each AEI to a low-dimensional feature subspace via the newly proposed two-dimensional locality preserving projections (2DLPP) method to further improve the discriminative power of the extracted features. Experimental results on the CASIA gait database (dataset B and C) demonstrate the effectiveness of the proposed method.

G0200-2-5 加速度・角速度センサーによる三次元歩行解析(バイオエンジニアリング部門一般講演(2),社会変革を技術で廻す機械工学)

Gait analysis is important for evaluation of walk ability and rehabilitation effect. We propose a new method for gait analysis using wearable acceleration and gyro sensors based on quaternion calculations. In this method, each subject fixed seven sensor units which consisted of a tri-axis acceleration sensor and three single axis gyro sensors on their body segments. Acceleration and angular velocity data detected from each sensor unit were corrected during level walking. Initial orientations of sensor units were shown as inclinations from gravity axis by using acceleration data during upright positioning in the trial. Angular displacements of sensor units were estimated using angular velocity data during through the trial. An original algorithm based on quaternion calculation was used for the orientation estimations. Each orientation of sensor unit was transformed to orientation of body segment by rotation matrix obtained from calibration trial. Body segment orientations were used for consisting 3-D stick animation of subject during trial for gait representation. Measurement and analysis were applied for five subjects. The results were compared with results from camera system to focus on the joint angle. The joint angles showed the same tendency with similar joint angle values obtained in both systems. The portability without spatial constraints of this sensor system could improve the capability for clinical applications based on the gait analysis.