Cross View Gait Recognition Using Joint-Direct Linear Discriminant Analysis

Establishment of the Lunar Phase Morphological Classification for Cervical Spinal Canal

STUDY DESIGN

Retrospective clinical trial.

PURPOSE

To establish a morphological classification of the cervical spinal canal using its parameters.

OVERVIEW OF LITERATURE

Cervical spine computed tomography (CT) data of 200 healthy volunteers in 2 years were analyzed. The morphology of the spinal cord was also analyzed.

METHODS

The median sagittal diameter and transverse diameter of the spinal canal from C2 to C7 were measured on CT images. The ratio of the median sagittal diameter to the transverse diameter was calculated. Accordingly, the spinal canal shape of each segment was classified into four, and the specific criteria of lunar phase classification were determined through linear discriminant analysis based on the ratio of the median sagittal diameter to the transverse diameter. The inter-rater reliability of the classification was explored using Kappa coefficients. Finally, the morphology of the different segments of the cervical spinal canal in healthy volunteers was revised and compared.

RESULTS

According to the ratio of the median sagittal diameter and the transverse diameter of the cervical spinal canal, the lunar phase classification of the cervical bony spinal canal was determined as follows: full-moon >0.65, 0.55< convex-moon ≤0.65, 0.46≤ quarter-moon ≤0.55, and residual-moon <0.46. The Kappa values of C2-C7 were 0.851, 0.958, 0.823, 0.927, 0.793, and 0.946, and the Kappa value of all C2-C7 segments was 0.854 that mainly presented two forms of full-moon (76.5%) and convex-moon (23.0%). A quarter-moon spinal canal was mainly distributed in C3, C4, C5, and C6; a residual-moon spinal canal was mainly distributed in C4 and C5; and the morphological distribution of C4 and C5 were similar (p>0.05). The frequency of the spinal canal of the residual-moon type was the highest, and the full-moon (6.5%) and residual-moon (7.5%) types of C7 were rare.

CONCLUSIONS

The morphological classification of the cervical spinal canal was established to present anatomical variations. The classification showed good inter-rater reliability.

Computer Vision and Machine Learning-Based Gait Pattern Recognition for Flat Fall Prediction

BACKGROUND

Gait recognition has been applied in the prediction of the probability of elderly flat ground fall, functional evaluation during rehabilitation, and the training of patients with lower extremity motor dysfunction. Gait distinguishing between seemingly similar kinematic patterns associated with different pathological entities is a challenge for the clinician. How to realize automatic identification and judgment of abnormal gait is a significant challenge in clinical practice. The long-term goal of our study is to develop a gait recognition computer vision system using artificial intelligence (AI) and machine learning (ML) computing. This study aims to find an optimal ML algorithm using computer vision techniques and measure variables from lower limbs to classify gait patterns in healthy people. The purpose of this study is to determine the feasibility of computer vision and machine learning (ML) computing in discriminating different gait patterns associated with flat-ground falls.

METHODS

We used the Kinect® Motion system to capture the spatiotemporal gait data from seven healthy subjects in three walking trials, including normal gait, pelvic-obliquity-gait, and knee-hyperextension-gait walking. Four different classification methods including convolutional neural network (CNN), support vector machine (SVM), K-nearest neighbors (KNN), and long short-term memory (LSTM) neural networks were used to automatically classify three gait patterns. Overall, 750 sets of data were collected, and the dataset was divided into 80% for algorithm training and 20% for evaluation.

RESULTS

The SVM and KNN had a higher accuracy than CNN and LSTM. The SVM (94.9 ± 3.36%) had the highest accuracy in the classification of gait patterns, followed by KNN (94.0 ± 4.22%). The accuracy of CNN was 87.6 ± 7.50% and that of LSTM 83.6 ± 5.35%.

CONCLUSIONS

This study revealed that the proposed AI machine learning (ML) techniques can be used to design gait biometric systems and machine vision for gait pattern recognition. Potentially, this method can be used to remotely evaluate elderly patients and help clinicians make decisions regarding disposition, follow-up, and treatment.

User Authentication by Gait Data from Smartphone Sensors Using Hybrid Deep Learning Network

User authentication and verification by gait data based on smartphones’ inertial sensors has gradually attracted increasing attention due to their compact size, portability and affordability. However, the existing approaches often require users to walk on a specific road at a normal walking speed to improve recognition accuracy. In order to recognize gaits under unconstrained conditions on where and how users walk, we proposed a Hybrid Deep Learning Network (HDLN), which combined the advantages of a long short-term memory (LSTM) network and a convolutional neural network (CNN) to reliably extract discriminative features from complex smartphone inertial data. The convergence layer of HDLN was optimized through a spatial pyramid pooling and attention mechanism. The former ensured that the gait features were extracted from more dimensions, and the latter ensured that only important gait information was processed while ignoring unimportant data. Furthermore, we developed an APP that can achieve real-time gait recognition. The experimental results showed that HDLN achieved better performance improvements than CNN, LSTM, DeepConvLSTM and CNN+LSTM by 1.9%, 2.8%, 2.0% and 1.3%, respectively. Furthermore, the experimental results indicated our model’s high scalability and strong suitability in real application scenes.

Recurrent Neural Network for Inertial Gait User Recognition in Smartphones

In this article, a gait recognition algorithm is presented based on the information obtained from inertial sensors embedded in a smartphone, in particular, the accelerometers and gyroscopes typically embedded on them. The algorithm processes the signal by extracting gait cycles, which are then fed into a Recurrent Neural Network (RNN) to generate feature vectors. To optimize the accuracy of this algorithm, we apply a random grid hyperparameter selection process followed by a hand-tuning method to reach the final hyperparameter configuration. The different configurations are tested on a public database with 744 users and compared with other algorithms that were previously tested on the same database. After reaching the best-performing configuration for our algorithm, we obtain an equal error rate (EER) of 11.48% when training with only 20% of the users. Even better, when using 70% of the users for training, that value drops to 7.55%. The system manages to improve on state-of-the-art methods, but we believe the algorithm could reach a significantly better performance if it was trained with more visits per user. With a large enough database with several visits per user, the algorithm could improve substantially.

Laplacian Eigenmaps Network-Based Nonlocal Means Method for MR Image Denoising

Magnetic resonance (MR) images are often corrupted by Rician noise which degrades the accuracy of image-based diagnosis tasks. The nonlocal means (NLM) method is a representative filter in denoising MR images due to its competitive denoising performance. However, the existing NLM methods usually exploit the gray-level information or hand-crafted features to evaluate the similarity between image patches, which is disadvantageous for preserving the image details while smoothing out noise. In this paper, an improved nonlocal means method is proposed for removing Rician noise in MR images by using the refined similarity measures. The proposed method firstly extracts the intrinsic features from the pre-denoised image using a shallow convolutional neural network named Laplacian eigenmaps network (LEPNet). Then, the extracted features are used for computing the similarity in the NLM method to produce the denoised image. Finally, the method noise of the denoised image is utilized to further improve the denoising performance. Specifically, the LEPNet model is composed of two cascaded convolutional layers and a nonlinear output layer, in which the Laplacian eigenmaps are employed to learn the filter bank in the convolutional layers and the Leaky Rectified Linear Unit activation function is used in the final output layer to output the nonlinear features. Due to the advantage of LEPNet in recovering the geometric structure of the manifold in the low-dimension space, the features extracted by this network can facilitate characterizing the self-similarity better than the existing NLM methods. Experiments have been performed on the BrainWeb phantom and the real images. Experimental results demonstrate that among several compared denoising methods, the proposed method can provide more effective noise removal and better details preservation in terms of human vision and such objective indexes as peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM).

A Novel Semi-Supervised Feature Extraction Method and Its Application in Automotive Assembly Fault Diagnosis Based on Vision Sensor Data

The fault diagnosis of dimensional variation plays an essential role in the production of an automotive body. However, it is difficult to identify faults based on small labeled sample data using traditional supervised learning methods. The present study proposed a novel feature extraction method named, semi-supervised complete kernel Fisher discriminant (SS-CKFDA), and a new fault diagnosis flow for automotive assembly was introduced based on this method. SS-CKFDA is a combination of traditional complete kernel Fisher discriminant (CKFDA) and semi-supervised learning. It adjusts the Fisher criterion with the data global structure extracted from large unlabeled samples. When the number of labeled samples is small, the global structure that exists in the measured data can effectively improve the extraction effects of the projected vector. The experimental results on Tennessee Eastman Process (TEP) data demonstrated that the proposed method can improve diagnostic performance, when compared to other Fisher discriminant algorithms. Finally, the experimental results on the optical coordinate data proves that the method can be applied in the automotive assembly process, and achieve a better performance.