Empowering lower limb disorder identification through PoseNet and artificial intelligence

Citation: Siddiqui, H.U.R.; Saleem, A.A.; Raza, M.A.; Villar, S.G.; Lopez, L.A.D.; Diez, I.d.l.T; Rustam, F.; Dudley, S. Empowering Lower Limb Disorder Identification through PoseNet and Artificial Intelligence. Diagnostics 2023,13, 2881. https://doi.org/10.3390/ diagnostics13182881 Academic Editor: Mehedi Masud Received: 15 July 2023 Revised: 25 August 2023 Accepted: 4 September 2023 Published: 8 September 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). diagnostics Article Empowering Lower Limb Disorder Identification through PoseNet and Artificial Intelligence Hafeez Ur Rehman Siddiqui 1, Adil Ali Saleem 1, Muhammad Amjad Raza 1, Santos Gracia Villar 2,3,4 , Luis Alonso Dzul Lopez 2,3,5 , Isabel de la Torre Diez 6,* , Furqan Rustam 7,* and Sandra Dudley 8 1Institute of Computer Science, Khwaja Fareed University of Engineering and Information Technology, Abu Dhabi Road, Rahim Yar Khan 64200, Punjab, Pakistan; [email protected] (H.U.R.S.); [email protected] (A.A.S.); [email protected] (M.A.R.) 2Universidad Europea del Atlántico, Isabel Torres 21, 39011 Santander, Spain; [email protected] (S.G.V.); [email protected] (L.A.D.L.) 3Universidad Internacional Iberoamericana, Campeche 24560, Mexico 4Department of Extension, Universidade Internacional do Cuanza, Cuito EN250, Bié, Angola 5Department of Project Management, Universidad Internacional Iberoamericana, Arecibo, PR 00613, USA 6Department of Signal Theory and Communications and Telematic Engineering, University of Valladolid, Paseo de Belén, 15, 47011 Valladolid, Spain 7School of Computer Science, University College Dublin, D04 V1W8 Dublin, Ireland 8Bioengineering Research Centre, School of Engineering, London South Bank University, 103 Borough Road, London SE1 0AA, UK; [email protected] *Correspondence: [email protected] (I.d.l.T.Z.); fur[email protected] (F.R.) Abstract: A novel approach is presented in this study for the classification of lower limb disorders, with a specific emphasis on the knee, hip, and ankle. The research employs gait analysis and the extraction of PoseNet features from video data in order to effectively identify and categorize these disorders. The PoseNet algorithm facilitates the extraction of key body joint movements and positions from videos in a non-invasive and user-friendly manner, thereby offering a comprehensive representation of lower limb movements. The features that are extracted are subsequently standardized and employed as inputs for a range of machine learning algorithms, such as Random Forest, Extra Tree Classifier, Multilayer Perceptron, Artificial Neural Networks, and Convolutional Neural Networks. The models undergo training and testing processes using a dataset consisting of 174 real patients and normal individuals collected at the Tehsil Headquarter Hospital Sadiq Abad. The evaluation of their performance is conducted through the utilization of K-fold cross-validation. The findings exhibit a notable level of accuracy and precision in the classification of various lower limb disorders. Notably, the Artificial Neural Networks model achieves the highest accuracy rate of 98.84%. The proposed methodology exhibits potential in enhancing the diagnosis and treatment planning of lower limb disorders. It presents a non-invasive and efficient method of analyzing gait patterns and identifying particular conditions. Keywords: lower limb disorder; PoseNet; gait analysis; machine learning; Artificial Neural Networks 1. Introduction Lower extremity disorders have been identified as a significant factor contributing to disability and reduced quality of life on a global scale [ 1 , 2 ]. Osteoarthritis of the knee, hip, and ankle is a commonly observed disorder affecting the lower limbs [ 3 , 4 ]. These conditions, commonly resulting from trauma, degenerative diseases, or bio-mechanical abnormalities, can give rise to discomfort, a restricted range of motion, and diminished functionality [ 5 , 6 ]. The prompt and precise classification of these conditions is crucial for effective treatment planning, individualized rehabilitation, and the prevention of further consequences. Clinical trials, subjective patient testimonials, and diagnostic imaging modalities such as X-rays and magnetic resonance imaging (MRI) have conventionally Diagnostics 2023,13, 2881. https://doi.org/10.3390/diagnostics13182881 https://www.mdpi.com/journal/diagnostics Diagnostics 2023,13, 2881 2 of 18 served as the predominant approaches in ascertaining the existence and extent of lower limb issues [ 7 , 8 ]. Although these techniques have proven to be valuable, they often require the utilization of specialized equipment, entail significant time investments [ 9 , 10 ], and may not fully capture the comprehensive dynamics of joint movement observed during typical activities. In recent years, advancements in technology have enabled the objective and continuous monitoring of the bio-mechanics in the lower extremities using gait data [ 11 , 12 ]. Gait analysis is employed in a wide variety of domains, including medical diagnostics [ 13 – 15 ], osteopathic medicine [ 16 , 17 ], comparative bio-mechanics [ 18 – 20 ], and sportsrelated bio-mechanics [ 21 – 23 ]. The application of gait analysis has exhibited considerable promise in the identification and assessment of lower limb disorders [ 24 ]. Gait analysis encompasses the evaluation of an individual’s walking pattern, including diverse elements such as stride length, step width, joint angles, and the temporal coordination of movements. Through the examination of these gait parameters, medical professionals and scholars can detect irregularities or deviations from typical gait patterns, which may serve as indicators of the existence of a lower limb disorder [ 25 – 27 ]. The objective of the study presented in this manuscript is to investigate the potential application of PoseNet features in the classification of lower limb disorders. PoseNet is a real-time pose estimation model developed by Google, which employs deep learning techniques to accurately assess human poses from both photos and videos [ 28 , 29 ]. Refs. [ 30 – 32 ] used PoseNet for in-home rehabilitation, and [ 33 , 34 ] used PoseNet for batsman stroke prediction. These features are capable of capturing the spatial positions and movements of key body joints. PoseNet was chosen by the authors because of its better attributes in the domain of real-time human pose estimation. This is due to its real-time capabilities, user friendliness, and versatility. The model’s ability to perform efficiently with just one video feed, together with its minimal preprocessing requirements and smooth integration into widely used deep learning frameworks, makes it an excellent choice for a wide range of applications. The researchers aimed to establish a dependable and precise classification system for the identification of particular disorders in the hip, ankle, and knee by means of analyzing gait patterns and movements extracted from videos. PoseNet offers a non-invasive and user-friendly approach to extracting human pose data from videos, obviating the necessity for dedicated apparatus or body-attached markers. This facilitates a more authentic and unimpeded evaluation of gait patterns in real-world contexts. The application of deep learning in PoseNet enables the extraction of complicated and detailed features from videos, resulting in a comprehensive depiction of movements in the lower limbs. The characteristics effectively capture the complex variations and fluctuations in an individual’s gait, which have the potential to serve as indicators for particular disorders affecting the lower limbs. This methodology presents the potential for improved accuracy in diagnosis. The study has the following contributions. • The data were collected from a total of 174 real patients and normal individuals, comprising both male and female participants. The data collection process involved capturing videos of the participants using a camera while they walked on a designated walkway at the Tehsil Headquarter (THQ) Hospital in Sadiqabad. • The data were gathered by the system via video recordings, thereby obviating the necessity for intrusive sensors or apparatus affixed to the subjects’ bodies. The implementation of this data collection method that minimizes interference guarantees a more authentic and unrestrained evaluation of gait patterns, thereby enhancing the ecological validity of the system. • The system employs PoseNet, a deep learning model, to extract relevant features from videos that capture movements of the lower limbs. By utilizing the features of PoseNet, the system capitalizes on the model’s capacity to accurately estimate the human pose, facilitating a thorough examination of gait patterns. • Through the application of machine learning (ML) algorithms to the extracted PoseNet features, the system possesses the capability to effectively classify and distinguish various disorders that impact the hip, ankle, and knee. The implementation of au- Diagnostics 2023,13, 2881 3 of 18 tomation in this context serves to decrease the level of subjectivity involved in manual analysis, while also reducing the amount of time required for such analysis. As a result, the process of diagnosis becomes more expedient and efficient. The aforementioned contributions play a significant role in the advancement of lower limb disorder classification, thereby holding the potential to yield substantial benefits for both clinical practice and research endeavors. The subsequent sections of the article are structured as follows: Section 2provides a comprehensive literature review, Section 3outlines the methodology and experimental procedures, and the obtained results are discussed in Section 4. Finally, the conclusions are presented in Section 5. 2. Literature Review In recent years, there has been growing interest in the field of gait analysis and the categorization of joint abnormalities. This attention is driven by the desire to enhance the accuracy of diagnosis and treatment methods. Various studies employ ML and deep learning techniques to automatically categorize joint abnormalities using gait data. Each study in the field of gait analysis is centered around a particular condition or aspect and employs a range of methodologies and evaluation metrics. The research conducted by [ 35 ] centers on the diagnosis of knee osteoarthritis using the automated analysis of walking data obtained from both diagnosed persons and symptom-free controls. Ground reaction force features are extracted using force plates and piezoelectric sensors, and these values are then associated with the severity of osteoarthritis using random forest regression models. The attained accuracy of 72.61% in the five-fold cross-validation indicates a decent level of performance, leaving space for potential improvement. In [ 36 ], the researchers use supervised classifiers and an RGB-D camera to diagnose gait problems in osteoarthritis patients. The researchers categorize gait disorders with 97% accuracy using fourteen gait measures, demonstrating its potential for osteoarthritis diagnosis. Another work [ 37 ] proposes a novel method of detecting gait abnormalities using a single 2D video camera. Video analysis with a support vector machine (SVM) classifier determines biomechanical gait parameters with 98.8% accuracy. The research in [ 38 ] presents a cost-effective and user-friendly gait data acquisition and analysis system. This technique quantifies osteoarthritis-related walking irregularities. The hybrid prediction model, combining manual and automated characteristics, achieves 98.77% accuracy. Meanwhile, Ref. [ 39 ] uses deep learning to classify abnormal gait patterns by integrating 3D skeletal data and plantar foot pressure readings. The multimodal hybrid model achieves 97.60% accuracy by utilizing pressure and skeletal data effectively. The aim of [ 40 ] was to develop an automated framework for knee osteoarthritis (KOA) classification utilizing radiographic imaging and gait analysis, with a Kallgren- Lawrence grading system. A support vector machine and deep learning features from Inception-ResNet-v2 classified KOA based on gait and radiographic data, showing strong relationships between gait characteristics and radiological severity. The AUC varied from 0.93 to 0.97 for KL grades 0–4. Moreover, Ref. [ 41 ] intended to evaluate gait symmetry in unilateral ankle osteoarthrosis (AOA) patients and identify variables affecting postsurgery asymmetry. They compared 46 gait metrics in 10 healthy people with 10 AOA patients using 3D inertial sensors and pressure insoles. They found significant differences in 23 impacted-side and 20 non-impacted-side variables. In particular, 14 metrics exhibited differences during bilateral AOA patient comparisons, notably in the toe area, and in forefoot mobility during walking. In [ 42 ], the researchers use ground reaction force (GRF) measurements to automate the diagnosis of functional gait disorders (GDs). They evaluate GRF parameterization methods for GD identification and establish a reference for automatic classification. The study divides 279 GD patients and 161 healthy controls into hip, knee, ankle, and calcaneus impairment groups using GRF data. It tests GRF and PCA-based parameterization approaches. The evaluation of discriminative power uses linear discriminant analysis. The study clas- Diagnostics 2023,13, 2881 4 of 18 sifies normal walking patterns and multiclass GD categories. The study in [ 43 ] focuses on categorizing gait disorders, with an emphasis on ground reaction force (GRF) analysis. The study preprocesses GRF signals and extracts and selects features from the GaitRec and Gutenberg databases with data from gait problem patients and healthy participants. The K- nearest neighbor (KNN) model outperforms conventional machine learning approaches in four experimental schemes categorizing gait disorders. The study contrasts vertical and three-dimensional GRF, with the latter performing better. Meanwhile, Ref. [ 44 ] develops an automated, accurate knee osteoarthritis (KOA) diagnosis method. The study uses RQA, fuzzy entropy, and statistical analysis to analyze dynamical characteristics collected from gait kinematic data. Discriminant analysis on these characteristics evaluates shallow classifiers like SVM, KNN, NB, DT, and Adaboost. SVM distinguishes KOA patients and healthy controls with the maximum accuracy of 92.31% and 100%, proving its KOA diagnostic efficacy. Previous research studies provide evidence of the efficacy of ML and deep learning methodologies in the automated categorization of joint abnormalities using gait data. The utilization of diverse modalities, including RGB-D cameras, 2D video, and ground reaction force measurements, exemplifies the multifaceted nature of these methodologies. Nevertheless, certain studies demonstrate limitations in terms of moderate accuracy, the necessity for supplementary evaluation metrics, and comparatively limited sample sizes. This manuscript introduces a novel approach to categorizing lower limb disorders, focusing on ankle, knee, hip, and normal subjects. The proposed method involves the utilization of PoseNet features extracted from video data. The approach centers on utilizing PoseNet, a pose estimation model based on deep learning, to extract significant features from the video recordings. The primary objective of the proposed methodology is to improve the precision and effectiveness of diagnosing lower limb disorders through the utilization of video data. This approach capitalizes on the abundance of valuable information pertaining to subjects’ movements and joint positions that can be extracted from video recordings. The application of this methodology holds promise in assisting healthcare practitioners in the identification and classification of distinct lower limb disorders, thus facilitating the implementation of suitable treatment and rehabilitation approaches. 3. Materials and Methods The methodology of this study was implemented on a Core i7 11th-generation machine running the Windows operating system. The proposed methodology was implemented using Jupyter Notebook [ 45 ], a web-based interactive tool that integrates live code execution, visualizations, and explanatory text. The Python programming language was used for the implementation. Several libraries, including sci-kit-learn and pandas, were utilized during the implementation process. Scikit-learn (sklearn) is a popular open-source machine learning library for Python that provides a comprehensive toolkit for classification, regression, and clustering [ 46 ]. Pandas is a well-known Python library for high-performance data analysis, with intuitive structures such as Series and DataFrame for structured data tasks such as cleansing, transformation, and aggregation [47]. 3.1. Proposed Methodology The research methodology employed for the classification of lower limb disorders is depicted in Figure 1. The data for this study were gathered at the Tehsil Headquarter (THQ) Hospital in Sadiq Abad through the application of a video camera. Then, the dataset was divided into training and testing sets. The frames were extracted from the videos to capture discrete temporal instances that were pertinent to the movement of the lower limbs. Subsequently, relevant features were derived from every frame utilizing the PoseNet algorithm. After performing feature extraction, the obtained features underwent standardization using a technique known as feature scaling. The standardized features were employed as inputs for a variety of ML algorithms. The algorithms underwent Diagnostics 2023,13, 2881 5 of 18 training and testing processes using a dataset to assess their efficacy in classifying various lower limb disorders, with a specific focus on hip, knee, and ankle ailments. Raw Videos Feature Extraction Data Data Splitting Testing Set Training Set Model Training Trained Model Prediction Evaluation Video Camera Moving Person Data Scaling Feature Reduction Figure 1. Proposed methodology diagram. 3.2. Data Collection The research was carried out in compliance with ethical protocols and received approval from the ethics committee at Khwaja Fareed University of Engineering and Technology (KFUEIT). The committee conducted a thorough evaluation of the ethical implications of the study, guaranteeing adherence to the principles delineated in the Helsinki Declaration. Through the acquisition of ethical approval, the study effectively showcased its dedication to safeguarding the well-being, entitlements, and confidentiality of the individuals engaged in the research. The data collection was conducted at the THQ Hospital located in Sadiq Abad. The data were obtained from a group of 174 individuals, ranging in age from 40 to 60 years, encompassing both male and female participants. Figure 2presents information pertaining to the number of participants and their corresponding genders. During the experiment, the participants were instructed to walk on a pathway measuring 9 m in length and data were recorded using a camera, as illustrated in Figure 3. The Hiievpu 2K Webcam was used for data collection as it exhibits sophisticated functionalities, effectively using the CMOS 1/s image sensor’s capability to provide high-definition visuals at a resolution of 4 million pixels (2560 × 1440p) [ 48 ]. This produces videos that are incredibly clear and operate smoothly at a frame rate of 30 frames per second [ 48 ]. To optimize patient comfort and convenience, all participants were given instructions to engage in a series of 10 walking sessions, with designated intervals of rest interspersed throughout. The objective of this approach was to acquire reliable data for subsequent processing, which involved extracting features and conducting further analysis using a video camera. Diagnostics 2023,13, 2881 6 of 18 Figure 2. Distribution of subjects and genders. Figure 3. Subject walking on the walkway in front of the camera. 3.3. Feature Extraction To achieve an accurate depiction of the lower limbs’ temporal features, frames were extracted from the videos at a rate of 30 frames per second (fps). The selection of this higher frame rate was made to record the motions linked to lower limb disorders precisely. Subsequently, the frames underwent the application of the PoseNet feature extraction technique. The PoseNet algorithm is extensively employed in the estimation of the human body pose [ 49 ]. The system examines the bodily posture in every frame and identifies significant landmarks, including joint positions, angles, and body key points. These landmarks offer valuable insights into the kinetics and spatial organization of the human body during walking. This study employed the Google MediaPipe library [ 50 ]; MediaPipe is a Google-developed open-source library that provides tools for the creation of perception pipelines that can process and analyze various forms of media data, such as images and Diagnostics 2023,13, 2881 7 of 18 videos, to extract information such as facial landmarks, hand tracking, and pose estimation. Here, the MediaPipe library was used to extract pose key points [ 28 ]. The MediaPipe library extracts 33 landmarks. In this research, only 22 distinct anatomical landmarks on the human body were extracted. The selection of 22 landmarks from the total of 33 was due to the exclusion of 11 head landmarks. The chosen landmarks represented the upper and lower limbs, as these segments are actively engaged during gait. The decision to include upper limb segment landmarks was influenced by the recognition of their roles beyond walking mechanics [ 51 , 52 ]. The movement of the arms and hands not only serves as a functional component of gait but also plays an essential role in maintaining balance [ 53 – 55 ]. During movement, the dynamic interplay between both the upper and lower extremities contributes to the overall stability and coordination of the human body. The landmarks extracted using the MediaPipe library are as follows. • Upper Body: –Left Shoulder; –Right Shoulder; –Left Elbow; –Right Elbow; –Left Wrist; –Right Wrist; –Left Pinky; –Right Pinky; –Left Index; –Right Index; –Left Thumb; –Right Thumb. • Lower Body: –Left Hip; –Right Hip; –Left Knee; –Right Knee; –Left Ankle; –Right Ankle; –Left Heel; –Right Heel; –Left Foot Index; –Right Foot Index. The mentioned landmarks were utilized as points of reference to denote the positioning of individuals within the frames of the video. The extraction of pose landmarks yielded significant data pertaining to the spatial positioning and arrangement of different body parts, such as the shoulders, elbows, wrists, hips, knees, and ankles. The accurate identification and analysis of joints and key points obtained through pose estimation play a vital role in effectively characterizing the posture and movements exhibited by individuals while walking. A widely employed method of representing pose landmarks involves the utilization of a skeletal representation. This process entails establishing connections between the pose landmarks through the use of lines, resulting in the formation of a skeletal framework that visually portrays the physical structure of the person, as illustrated in Figure 4. Only blue landmarks were selected in this study. Through the integration of key characteristics, a concise yet informative anatomical framework was established, showcasing the overall physical posture and organization. The X, Y, and Z coordinates of these pivotal points, along with their corresponding visibility scores, were derived from the Cartesian coordinate system [ 28 ]. In order to maintain the integrity of the data, they were subsequently stored in a Comma-Separated Values (CSV) file. The structure of the CSV file was such that each row corresponded to an individual instance or frame Diagnostics 2023,13, 2881 8 of 18 of the video. The columns within the file represented the X, Y, and Z coordinates of the pose landmarks, along with their respective visibility scores. This format guaranteed the systematic arrangement and convenient accessibility of the crucial data about the 22 pose landmarks, facilitating their subsequent analysis and processing. • Number of key points = 22. • Number of extracted features from each key point = 4. • Total number of features = 22 × 4 = 88. 3.4. Data Scaling and Feature Reduction Feature scaling plays a pivotal role in the data preprocessing phase that follows feature extraction. The primary objective of this process is to ensure that all extracted features undergo a uniform transformation in relation to their scale or range. This step holds significance due to the potential for significant variances in value ranges among different features, which can impact the performance of specific ML algorithms. Feature scaling is a technique that mitigates the disproportionate influence of features with larger magnitudes on the model training process. The standard scaler technique was utilized in this manuscript for feature scaling [ 56 ]. The standard scaler is a widely employed technique for the normalization of feature values. The process involves centering the feature values by subtracting the mean and scaling them by dividing them by the standard deviation, resulting in a mean of zero and a standard deviation of one. The standard scaler accomplishes this by performing two operations on each data point: subtracting the mean value of the feature and dividing the result by the standard deviation. The equation for the standard scaler can be defined as Xscaled = X−µ σ(1) where •Xscaled represents the scaled feature value; •Xis the original feature value; •µis the mean of the feature values in the dataset; •σis the standard deviation of the feature values in the dataset. Figure 4. Selected landmarks of subject in blue. Diagnostics 2023,13, 2881 9 of 18 The normalization process guarantees that the feature values are standardized, irrespective of their initial distribution, by centering them around zero and applying uniform scaling. Following the application of the standard scaler to the dataset, the subsequent step in the methodology entailed the implementation of Principal Component Analysis (PCA) on the scaled features. PCA is a statistical methodology employed to achieve dimensionality reduction [ 57 ]. The main aim of this approach is to identify the most prominent patterns or variations within the dataset by converting the initial features into a distinct set of independent variables known as principal components. The dataset was represented by a selection of 80 principal components in this study. The selection of these components was based on their capacity to capture the greatest proportion of variance in the data while minimizing the loss of information. The principal components obtained were utilized for the training of different ML models. 3.5. Exploratory Data Analysis The dataset consists of a heterogeneous compilation of 1740 videos, encompassing a range of categories including hip, ankle, knee, and normal. Figure 5, through informative pie charts, provides a visual representation of the distribution of videos in the training and testing sets among the different categories. The charts not only display the number of videos within each category but also emphasize the relative significance of each category by presenting their percentage representation in the overall dataset. (a) (b) Figure 5. Distribution of videos in each category: (a) training set (b) testing set. Upon analyzing Figure 5a,b, it becomes evident that there is a discernible pattern in the composition of the dataset. Specifically, the normal category displays a substantially greater number of contributions in comparison to the knee, ankle, and hip categories. The distribution of categories within the dataset offers valuable insights into their prevalence, enabling a comprehensive understanding of the dataset’s composition and its implications for both research and practical applications. The number of frames extracted from each category is shown in Figure 6. Diagnostics 2023,13, 2881 16 of 18 S.D.: Conceptualization, Formal Analysis, Funding Acquisition, Project Administration, Resources, Supervision, Validation, Writing—Review and Editing. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of KFUEIT (HREC- 22-692 approved on 20 June 2022). Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: Data will be provided on demand. Conflicts of Interest: The authors declare no conflict of interest. Abbreviations The following abbreviations are used in this manuscript: RF Random Forest DT Decision Tree KNN K-Nearest Neighbor LR Logistic Regression SVM Support Vector Machine PCA Principal Component Analysis VNC Virtual Network Computing FPS Frames Per Second NN Neural Network CNN Convolutional Neural Network ML Machine Learning AOA Ankle Osteoarthrosis ETC Extra Tree Classifier GDs Gait Disorders ELA Learning-Based Adaboost RQA Recurrence Quantification Analysis References 1. Fatima, S.Z. Life of an amputee: Predictors of quality of life after lower limb amputation. Wien. Med. Wochenschr. 2022 , 1–5. [CrossRef] [PubMed] 2. Grimmer, M.; Riener, R.; Walsh, C.J.; Seyfarth, A. Mobility related physical and functional losses due to aging and disease-a motivation for lower limb exoskeletons. J. Neuroeng. Rehabil. 2019,16, 1–21. [CrossRef] 3. Pirani, H.; Noori, S.; Shahmoradi, D. Examining the prevalence of lower body disorders among male lower-secondary education students in Kermanshah. Int. J. Health Life Sci. 2019,5, e85033. [CrossRef] 4. Leggit, J.; Carey, P.M.; Alisangco, J.B., Disorders of the Lower Extremity. In Family Medicine: Principles and Practice; Paulman, P.M., Taylor, R.B., Paulman, A.A., Nasir, L.S., Eds.; Springer International Publishing: Cham, Switzerland, 2017; pp. 1489–1513. [CrossRef] 5. Osteoarthritis. Available online: https://www.mayoclinic.org/diseases-conditions/osteoarthritis/symptoms-causes/syc-203519 25 (accessed on 17 August 2023). 6. Musculoskeletal Health. 2022. Available online: https://www.who.int/news-room/fact-sheets/detail/musculoskeletalconditions (accessed on 17 August 2023). 7. Khalid, H.; Hussain, M.; Al Ghamdi, M.A.; Khalid, T.; Khalid, K.; Khan, M.A.; Fatima, K.; Masood, K.; Almotiri, S.H.; Farooq, M.S.; et al. A comparative systematic literature review on knee bone reports from mri, x-rays and ct scans using deep learning and machine learning methodologies. Diagnostics 2020,10, 518. [CrossRef] [PubMed] 8. Heidari, B. Knee osteoarthritis diagnosis, treatment and associated factors of progression: Part II. Casp. J. Intern. Med. 2011 ,2, 249. 9. Hamza, A.; Khan, M.A.; Alhaisoni, M.; Al Hejaili, A.; Shaban, K.A.; Alsubai, S.; Alasiry, A.; Marzougui, M. D2BOF-COVIDNet: A framework of deep bayesian optimization and fusion-assisted optimal deep features for COVID-19 classification using chest X-ray and mri scans. Diagnostics 2022,13, 101. [CrossRef] 10. Devereux, R.B.; Pini, R.; Aurigemma, G.P.; Roman, M.J. Measurement of left ventricular mass: Methodology and expertise. J. Hypertens. 1997,15, 801–809. [CrossRef] Diagnostics 2023,13, 2881 17 of 18 11. Weygers, I.; Kok, M.; Konings, M.; Hallez, H.; De Vroey, H.; Claeys, K. Inertial sensor-based lower limb joint kinematics: A methodological systematic review. Sensors 2020,20, 673. [CrossRef] 12. Picerno, P. 25 years of lower limb joint kinematics by using inertial and magnetic sensors: A review of methodological approaches. Gait Posture 2017,51, 239–246. [CrossRef] 13. Balaji, E.; Brindha, D.; Elumalai, V.K.; Umesh, K. Data-driven gait analysis for diagnosis and severity rating of Parkinson’s disease. Med Eng. Phys. 2021,91, 54–64. 14. Van Deventer, K.A.; Seehusen, C.N.; Walker, G.A.; Wilson, J.C.; Howell, D.R. The diagnostic and prognostic utility of the dual-task tandem gait test for pediatric concussion. J. Sport Health Sci. 2021,10, 131–137. [CrossRef] [PubMed] 15. Beyrami, S.M.G.; Ghaderyan, P. A robust, cost-effective and non-invasive computer-aided method for diagnosis three types of neurodegenerative diseases with gait signal analysis. Measurement 2020,156, 107579. [CrossRef] 16. Hill, C.N.; Romero, M.; Rogers, M.; Queen, R.M.; Brolinson, P.G. Effect of osteopathic manipulation on gait asymmetry. J. Osteopath. Med. 2021,122, 85–94. [CrossRef] 17. Terrell, Z.T.; Moudy, S.C.; Hensel, K.L.; Patterson, R.M. Effects of osteopathic manipulative treatment vs. osteopathic cranial manipulative medicine on Parkinsonian gait. J. Osteopath. Med. 2022,122, 243–251. [CrossRef] [PubMed] 18. Martins, J.S.; Sabino, G.; Nascimento, D.H.; Machado, G.M.; Vimieiro, C.B. Biomechanical model for Dynamic Analysis of the Human Gait. In Proceedings of the International Symposium on Computer Methods in Biomechanics and Biomedical Engineering, New York, NY, USA, 14–16 August 2019; pp. 362–370. 19. Silva, M.P.; Ambrósio, J.A. Sensitivity of the results produced by the inverse dynamic analysis of a human stride to perturbed input data. Gait Posture 2004,19, 35–49. [CrossRef] [PubMed] 20. Ruiz, D.V.; Magluta, C.; Roitman, N. Experimental verification of biomechanical model of bipedal walking to simulate vertical loads induced by humans. Mech. Syst. Signal Process. 2022,167, 108513. [CrossRef] 21. Song, Y.; Biro, I. The Evolution of Marker-based Motion Analysis and the Integration of Advanced Computational Methods: Application to Human Gait Biomechanics. In Proceedings of the 2022 2nd International Conference on Bioinformatics and Intelligent Computing, Harbin, China, 21–23 January 2022; pp. 201–205. 22. Li, J.; Du, H. Research on the sports biomechanics modeling of the human motion technical movements. In Cyber Security Intelligence and Analytics; Springer: Berlin/Heidelberg, Germany, 2020; pp. 243–248. 23. Meng, Y.; Bíró, I.; Sárosi, J. Markerless measurement techniques for motion analysis in sports science. Analecta Tech. Szeged. 2023 , 17, 24–31. [CrossRef] 24. Fouasson-Chailloux, A.; Menu, P.; Dauty, M. Lower-Limb Arthropathies and Walking: The Use of 3D Gait Analysis as a Relevant Tool in Clinical Practice. Int. J. Environ. Res. Public Health 2022,19, 6785. [CrossRef] 25. Klöpfer-Krämer, I.T.; Augat, P., Functional Capacity Evaluation and Quantitative Gait Analysis: Lower Limb Disorders. In Handbook of Human Motion; Müller, B., Wolf, S.I., Brueggemann, G.P., Deng, Z., McIntosh, A., Miller, F., Selbie, W.S., Eds.; Springer International Publishing: Cham, Switzerland, 2017; pp. 1–17. [CrossRef] 26. Shin, J.; Yoo, J.; Jung, T.; Goo, H. Comparison of lower extremity motor score parameters for patients with motor incomplete spinal cord injury using gait parameters. Spinal Cord 2011,49, 529–533. [CrossRef] 27. Schmid-Zalaudek, K.; Fischer, T.; Száva, Z.; Lackner, H.K.; Kropiunig, U.; Bittner, C.; Höcker, K.; Winkler, G.; Peternell, G. Kinetic Gait Parameters in Unilateral Lower Limb Amputations and Normal Gait in Able-Bodied: Reference Values for Clinical Application. J. Clin. Med. 2022,11, 2683. [CrossRef] 28. Fernandez, J. MediaPipe Pose Estimation Documentation. Available online: https://github.com/google/mediapipe/blob/ master/docs/solutions/pose.md (accessed on 17 August 2023). 29. Chen, Y.; Shen, C.; Wei, X.S.; Liu, L.; Yang, J. Adversarial posenet: A structure-aware convolutional network for human pose estimation. In Proceedings of the IEEE International Conference on Computer Vision, Venice, Italy, 22–29 October 2017; pp. 1212–1221. 30. Chua, J.; Ong, L.Y.; Leow, M.C. Telehealth using PoseNet-based system for in-home rehabilitation. Future Internet 2021 ,13, 173. [CrossRef] 31. Li, Y.; Wang, C.; Cao, Y.; Liu, B.; Tan, J.; Luo, Y. Human pose estimation based in-home lower body rehabilitation system. In Proceedings of the 2020 International Joint Conference on Neural Networks (IJCNN), Glasgow, UK, 19–24 July 2020; pp. 1–8. 32. Cordella, F.; Di Corato, F.; Zollo, L.; Siciliano, B. A robust hand pose estimation algorithm for hand rehabilitation. In Proceedings of the New Trends in Image Analysis and Processing–ICIAP 2013: ICIAP 2013 International Workshops, Naples, Italy, 9–13 September 2013; pp. 1–10. 33. Siddiqui, H.U.R.; Younas, F.; Rustam, F.; Flores, E.S.; Ballester, J.B.; Diez, I.d.l.T.; Dudley, S.; Ashraf, I. Enhancing Cricket Performance Analysis with Human Pose Estimation and Machine Learning. Sensors 2023,23, 6839. [CrossRef] [PubMed] 34. Devanandan, M.; Rasaratnam, V.; Anbalagan, M.K.; Asokan, N.; Panchendrarajan, R.; Tharmaseelan, J. Cricket Shot Image Classification Using Random Forest. In Proceedings of the 2021 3rd International Conference on Advancements in Computing (ICAC), Colombo, Sri Lanka, 9–11 December 2021; pp. 425–430. 35. Kotti, M.; Duffell, L.D.; Faisal, A.A.; McGregor, A.H. Detecting knee osteoarthritis and its discriminating parameters using random forests. Med. Eng. Phys. 2017,43, 19–29. [CrossRef] 36. Cui, X.; Zhao, Z.; Ma, C.; Chen, F.; Liao, H. A gait character analyzing system for osteoarthritis pre-diagnosis using RGB-D camera and supervised classifier. In Proceedings of the World Congress on Medical Physics and Biomedical Engineering 2018, Prague, Czech Republic, 3–8 June 2018; Springer: Berlin/Heidelberg, Germany, 2019; Volume 1, pp. 297–301. Diagnostics 2023,13, 2881 18 of 18 37. Verlekar, T.T.; Soares, L.D.; Correia, P.L. Automatic classification of gait impairments using a markerless 2D video-based system. Sensors 2018,18, 2743. [CrossRef] [PubMed] 38. Chen, F.; Cui, X.; Zhao, Z.; Zhang, D.; Ma, C.; Zhang, X.; Liao, H. Gait acquisition and analysis system for osteoarthritis based on hybrid prediction model. Comput. Med Imaging Graph. 2020,85, 101782. [CrossRef] 39. Jun, K.; Lee, S.; Lee, D.W.; Kim, M.S. Deep learning-based multimodal abnormal gait classification using a 3D skeleton and plantar foot pressure. IEEE Access 2021,9, 161576–161589. [CrossRef] 40. Kwon, S.B.; Han, H.S.; Lee, M.C.; Kim, H.C.; Ku, Y. Machine learning-based automatic classification of knee osteoarthritis severity using gait data and radiographic images. IEEE Access 2020,8, 120597–120603. [CrossRef] 41. Chopra, S.; Crevoisier, X. Preoperative gait asymmetry in end-stage unilateral ankle osteoarthrosis patients. Foot Ankle Surg. 2019,25, 298–302. [CrossRef] 42. Slijepcevic, D.; Zeppelzauer, M.; Gorgas, A.M.; Schwab, C.; Schüller, M.; Baca, A.; Breiteneder, C.; Horsak, B. Automatic classification of functional gait disorders. IEEE J. Biomed. Health Inform. 2017,22, 1653–1661. [CrossRef] 43. Shuzan, M.N.I.; Chowdhury, M.E.; Reaz, M.B.I.; Khandakar, A.; Abir, F.F.; Faisal, M.A.A.; Ali, S.H.M.; Bakar, A.A.A.; Chowdhury, M.H .; Mahbub, Z.B.; et al. Machine learning-based classification of healthy and impaired gaits using 3D-GRF signals. Biomed. Signal Process. Control 2023,81, 104448. [CrossRef] 44. Zeng, W.; Ma, L.; Zhang, Y. Detection of knee osteoarthritis based on recurrence quantification analysis, fuzzy entropy and shallow classifiers. Multimed. Tools Appl. 2023, 1–22. [CrossRef] 45. Jupyter Notebook. Available online: https://jupyter.org/ (accessed on 17 August 2023). 46. scikit-learn. Available online: https://scikit-learn.org/stable/index.html (accessed on 17 August 2023). 47. Pandas. Available online: https://pandas.pydata.org/ (accessed on 17 August 2023). 48. 2K QHD Webcam. Available online: https://hiievpu.us/products/2k-qhd-webcam?pr_prod_strat=copurchase_transfer_ learning&pr_rec_id=39d1bbc4a&pr_rec_pid=7619743645942&pr_ref_pid=7619750461686&pr_seq=uniform (accessed on 17 August 2023). 49. Fei, K.; Wang, C.; Zhang, J.; Liu, Y.; Xie, X.; Tu, Z. Flow-pose Net: An effective two-stream network for fall detection. Vis. Comput. 2022,39, 2305–2320. [CrossRef] 50. MediaPipe Solutions Guide. Available online: https://developers.google.com/mediapipe/solutions/guide (accessed on 17 August 2023). 51. Meyns, P. Comment: Difference between assessment of upper limb movement and upper limb associated reactions during walking. J. NeuroEng. Rehabil. 2021,18, 47. [CrossRef] [PubMed] 52. Stephenson, J.L.; Lamontagne, A.; De Serres, S.J. The coordination of upper and lower limb movements during gait in healthy and stroke individuals. Gait Posture 2009,29, 11–16. [CrossRef] [PubMed] 53. Bruijn, S.M.; Meijer, O.G.; Beek, P.J.; Van Dieen, J.H. The effects of arm swing on human gait stability. J. Exp. Biol. 2010 , 213, 3945–3952. [CrossRef] 54. Ferraris, C.; Amprimo, G.; Masi, G.; Vismara, L.; Cremascoli, R.; Sinagra, S.; Pettiti, G.; Mauro, A.; Priano, L. Evaluation of Arm Swing Features and Asymmetry during Gait in Parkinson’s Disease Using the Azure Kinect Sensor. Sensors 2022,22, 6282. [CrossRef] 55. Hill, A.; Nantel, J. The effects of arm swing amplitude and lower-limb asymmetry on gait stability. PLoS ONE 2019 ,14, e0218644. [CrossRef] 56. Sharma, V. A Study on Data Scaling Methods for Machine Learning. Int. J. Glob. Acad. Sci. Res. 2022,1, 23–33. [CrossRef] 57. Ivosev, G.; Burton, L.; Bonner, R. Dimensionality reduction and visualization in principal component analysis. Anal. Chem. 2008 , 80, 4933–4944. [CrossRef] 58. Ahsan, M.M.; Luna, S.A.; Siddique, Z. Machine-learning-based disease diagnosis: A comprehensive review. Healthcare 2022 , 10, 541. [CrossRef] 59. George, A.; Ravindran, A.; Mendieta, M.; Tabkhi, H. Mez: An adaptive messaging system for latency-sensitive multi-camera machine vision at the iot edge. IEEE Access 2021,9, 21457–21473. [CrossRef] 60. Mendieta, M.; Neff, C.; Lingerfelt, D.; Beam, C.; George, A.; Rogers, S.; Ravindran, A.; Tabkhi, H. A novel application/infrastructure co-design approach for real-time edge video analytics. In Proceedings of the 2019 SoutheastCon, Huntsville, AL, USA, 11–14 April 2019; pp. 1–7. 61. Sharma, A.; Jain, A.; Gupta, P.; Chowdary, V. Machine learning applications for precision agriculture: A comprehensive review. IEEE Access 2020,9, 4843–4873. [CrossRef] 62. Shafique, R.; Aljedaani, W.; Rustam, F.; Lee, E.; Mehmood, A.; Choi, G.S. Role of Artificial Intelligence in Online Education: A Systematic Mapping Study. IEEE Access 2023,11, 52570–52584. [CrossRef] 63. Brownlee, J. A Gentle Introduction to k-Fold Cross-Validation. 2020. Available online: https://machinelearningmastery.com/kfold-cross-validation/ (accessed on 17 August 2023). Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.