Fig 6 - uploaded by Aftab Khan
Content may be subject to copyright.
Illustration of the bowling lengths during our batting sessions. Some preferred shot types for those lengths are also illustrated. Decision regarding off and on side is usually made based on the line of the ball.
Source publication
Quality assessment in cricket is a complex task that is performed by understanding the combination of individual activities a player is able to perform and by assessing how well these activities are performed. We present a framework for inexpensive and accessible, automated recognition of cricketing shots. By means of body-worn inertial measurement...
Contexts in source publication
Context 1
... detailed breakdown of the collected dataset is given in Table 2. Each session was on average 15.63 ± 5.44 minutes long, where sensors (each equipped with an accelerometer, a gyroscope and a magnetometer) were placed on all four limbs of the players (as illustrated in Figure 6; note that sensors on the lower limbs were placed behind the protective pads). Due to the high speed motion of batting shots, sampling rate of 100Hz was used (higher than required for standard HAR models [23]). ...
Context 2
... length (point in front of the batsman where the ball is pitched) was varied between 'short pitched' and 'full pitched' (cf. Figure 6). ...
Context 3
... of the shots are usually played according to the type of the incoming ball (see Figure 6 for various types of bowling lengths for example). Shot analysis, therefore, can be performed in order to see how batsmen execute their shots in response to the kind of bowling they are exposed to. ...
Context 4
... only does this reflect the change in the configuration of the bowling machine during these sessions but also the change in strategy of shots by the player. In sessions 1 and 2, the bowling machine was kept at a full to good length (see Figure 6) and therefore straighter drives south of the ground are the most favoured and can clearly be observed. ...
Context 5
... bowling line and length play a major role in a batsman's shot choice of batting, as well as in the method the shot is played (see Figure 6). For example, a short length ball is more likely to be pulled, i.e., shot towards the on side of the ground played using a back-foot stance, whilst a full length ball is more likely to be driven (usually played straight down the ground). ...
Similar publications
Activity recognition is inherent to the vision of Internet-of-Things-enabled smart environments and enables end-users to perform activities of interest and have their smart environments respond appropriately. However as smart environments evolve and are expanded over time, the effort that end-users and others put into training their activity recogn...
In this paper we summarize the contributions of participants to the Sussex-Huawei Transportation-Locomotion (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp 2018. The SHL challenge is a machine learning and data science competition, which aims to recognize eight transportation activities (Still, Walk, Run, Bike, Bus, Car, Trai...
Human activity recognition (HAR) is an important research area due to its potential for building context-aware interactive systems. Though movement-based activity recognition is an established area of research, recognising sedentary activities remains an open research question. Previous works have explored eye-based activity recognition as a potent...
Labeled data used for training activity recognition classifiers are usually limited in terms of size and diversity. Thus, the learned model may not generalize well when used in real-world use cases. Semi-supervised learning augments labeled examples with unlabeled examples, often resulting in improved performance. However, the semi-supervised metho...
Citations
... However, similar studies published demonstrated successful applications with relatively small participant samples. Khan et al. [60] explored activity recognition in cricket using only six participants, mostly amateurs, highlighting the feasibility of such systems even with limited data. Hölzemann and Van Laerhoven [61] achieved promising results in recognizing basketball activities with IMUs using only three participants. ...
1) Background: Tennis has changed toward power-driven gameplay, demanding a nu-anced understanding of performance factors. This review explores the role of machine learning in enhancing tennis performance. (2) Methods: A systematic search identified articles utilizing machine learning in tennis performance analysis. (2) Results: Machine learning applications show promise in psychological state monitoring, talent identification, match outcome prediction, spatial and tactical analysis, and injury prevention. Coaches can leverage wearable technologies for per-sonalized psychological state monitoring, data-driven talent identification, and tactical insights for informed decision-making. (4) Conclusions: Machine learning offers coaches insights to refine coaching methodologies and optimize player performance in tennis. By integrating these insights, coaches can adapt to the demands of the sport by improving the players' outcomes. As technology progresses, continued exploration of machine learning's potential in tennis is warranted for further advancements in performance optimization.
... In tennis, multiclass classification was used to detect and classify stroke type (Brzostowski & Szwach, 2018;Whiteside, Cant, Connolly, & Reid, 2017), and shot type (Anand et al., 2017;Ganser, Hollaus, & Stabinger 2021). A CNN architecture was used to differentiate types of shooting postures in basketball (Fan, Bi, Wang, Zhang, & Sun, 2021), an SVM model was used to classify throw type and approach in handball (van den Tillaar, Bhandurge, & Stewart, 2021) and to analyze shots in cricket (Khan, Nicholson, & Plötz, 2017). ...
... Grade level was specified in cricket (McGrath, Neville, Stewart, & Cronin, 2019), Australian football (Cust, Sweeting, Ball, & Robertson, 2021), tennis (Whiteside et al., 2017) and soccer . Other studies reported the amateur professionalism in basketball (Hollaus et al., 2020), badminton (Steels et al., 2020), cricket (Jowitt, Durussel, Brandon, & King, 2020;Khan et al., 2017;McGrath, Neville, Stewart, Clinning, & Cronin, J, 2021), golf (Kim & Park, 2020), handball (van den Tillaar et al., 2021), netball (Smith & Bedford, 2020), tennis (Brzostowski & Szwach 2018;Ganser et al., 2021), volleyball (Kautz et al., 2017;Wang et al., 2018), and years of experience was reported in basketball (Fan et al., 2021). Wang, Guo, and Zhao (2016) specified that the badminton players were either members of the university club or team. ...
... Detection methods to segment the motion states were reported in 22 studies. Window segmentation which involves splitting the sensor signals into distinct windows of time (measured in seconds), was the most common method of activity detection (Bo, 2022;Brzostowski & Szwach, 2018;Cust et al., 2021;Jowitt et al., 2020;Khan et al., 2017;Kim and Park, 2020;McGrath et al., 2019;McGrath et al. 2021;Reilly et al., 2021;Salim et al., 2020;Salman et al., 2017;Shahar, Ghazali, As✁ari, & Swee, 2020). Smith & Bedford, 2020;Steels et al., 2020;Tan & Xie, 2021;van den Tillaar et al., 2021;Whiteside et al., 2017;Zhang et al., 2021). ...
There is an ever-present need to objectively measure and analyze sports motion for the determination of correct patterns of motion for skill execution. Developments in performance analysis technologies such as inertial measuring units (IMUs) have resulted in enormous data generation. However, these advances present challenges in analysis, interpretation, and transformation of data into useful information. Artificial intelligence (AI) systems can process and analyze large amounts of data quickly and efficiently through classification techniques. This study aimed to systematically review the literature on Machine Learning (ML) and Deep Learning (DL) methods applied to IMU data inputs for evaluating techniques or skills in individual swing and team sports. Electronic database searches (IEEE Xplore, PubMed, Scopus, and Google Scholar) were conducted and aligned with the PRISMA statement and guidelines. A total of 26 articles were included in the review. The Support Vector Machine (SVM) was identified as the most utilized model, as per 7 studies. A deep learning approach was reported in 6 studies, in the form of a Convolutional Neural Network (CNN) architecture. The in-depth analysis highlighted varying methodologies across all sports inclusive of device specifications, data preprocessing techniques and model performance evaluation. This review highlights that each step of the ML modeling process is iterative and should be based on the specific characteristics of the movement being analyzed.
... While other fields focus on reducing the number and size of sensors used in research, the sports industry is adopting multimodal sensors to gain a comprehensive understanding of player movements and physical states. These sensors facilitate various analyses, from simple posture classification to performance analysis, and include inertial measurement units (IMUs) [24][25][26][27][28][29] , eye trackers [30][31][32] , pressure sensors 33,34 , skeleton tracking sensors [35][36][37][38][39][40] , electromyography (EMG) sensors 41,42 , and capacitive sensors 43 . ...
The sports industry is witnessing an increasing trend of utilizing multiple synchronized sensors for player data collection, enabling personalized training systems with multi-perspective real-time feedback. Badminton could benefit from these various sensors, but there is a scarcity of comprehensive badminton action datasets for analysis and training feedback. Addressing this gap, this paper introduces a multi-sensor badminton dataset for forehand clear and backhand drive strokes, based on interviews with coaches for optimal usability. The dataset covers various skill levels, including beginners, intermediates, and experts, providing resources for understanding biomechanics across skill levels. It encompasses 7,763 badminton swing data from 25 players, featuring sensor data on eye tracking, body tracking, muscle signals, and foot pressure. The dataset also includes video recordings, detailed annotations on stroke type, skill level, sound, ball landing, and hitting location, as well as survey and interview data. We validated our dataset by applying a proof-of-concept machine learning model to all annotation data, demonstrating its comprehensive applicability in advanced badminton training and research.
... Nicholson et al. [15] have proposed an Activity recognition technology using Hierarchical Representation to assess the quality of cricket shots. Automated recognition of cricket shots is done using low-cost hardware and a simple setup that allows players to analyze the statistics of a shot played. ...
This book series aims to provide a forum for researchers from both academia and industry to share their latest research contributions in the area of computing technologies and Data Sciences and thus to exchange knowledge with the common goal of shaping the future. The best way to create memories is to gather and share ideas, creativity and innovations. The content of the book is as follows
... Regardless of the number of informational cues, or the quality of these, they must all be used by batters to produce a singular batting stroke. While there are over 25 different batting strokes (based on human annotations of stroke types commonly used in performance analysis), most of them are classified as front-foot or back-foot strokes based on the batter's gross body movement -whether the batter moves forward or backward from their original stance (Khan et al., 2017). These front-foot or back-foot strokes have been studied in relation to the biomechanical and kinematic determinants of individual strokes such as the front foot drive (Peploe et al., 2014;Stretch et al., 1998), but have not been studied together in an interactive, competitive environment where they serve as decision-driven actions. ...
... Based on the statistics of 366 matches, the study's accuracy was 71%. The study by Aftab Khan et al. (47) offers a framework for the automated recognition of cricket shots that is both affordable and practical. The movements of the batsmen are captured using body-worn inertial measurement units, and the data is subsequently analysed using a parallelized, hierarchical recognition system that automatically identifies pertinent categories of strokes as necessary for evaluating batting quality. ...
Introduction
Sports of all kinds even though have an alluring property of keeping their onlookers stuck to their place, the introduction of Technology, however, revolutionized it all together. Not only in legal sports but also the training and teaching methods have been reformed. The use of Information Communication and Technology (ICT) based technologies [Convolutional Neural Networks (CNN), Hawkeye, Computer vision, Artificial intelligence, etc.] has moderately increased the interactive nature of sports. Employing ICT-driven technologies have continuously been increasing performance levels due to which high effective performance levels have been achieved. In addition to offering information to the users, it also acts as a means for connecting and interacting with the remaining world. In this article, we provide a review of the studies considering the developments and impact of employing ICT technology on sports, especially cricket. The study has focussed on domain-specific developments in cricket sports: developments in the batting domain, bowling domain, and wicketkeeping as well.
Methods
For the study, the analysis has been done following the PRISMA guidelines.
Results
The study found that even though the researchers have done justifiable work in employing technology in sports as a whole but the domain-specific contribution in sports like cricket is not at the level as is need of the hour. In addition to the mentioned domains in the study, the research should gain speed in other domains like domain-specific Talent Identification for both genders, different age groups, diverse sports, etc.
Discussion
undoubtedly, the sports domain is employing technology at a vast level but a few domains like sports talent identification especially related to the most famous games like cricket require an immediate and intense focus of the researchers. Since this domain is still carrying out a traditional coach-oriented approach. There is an acute need to revolutionize the domain by incorporating modern technologies into it.
... The widespread availability of commodity wearables such as smartphones and smartwatches, has resulted in increased interest towards their utilization for applications such as sports and fitness tracking (e.g., running, bicycling, and swimming) [40,42,43,56]. These devices benefit from onboard sensors, including Inertial Measurement Units (IMUs), which can track and measure human movements that are subsequently analyzed for understanding activities. ...
Human activity recognition (HAR) in wearable computing is typically based on direct processing of sensor data. Sensor readings are translated into representations, either derived through dedicated preprocessing, or integrated into end-to-end learning. Independent of their origin, for the vast majority of contemporary HAR, those representations are typically continuous in nature. That has not always been the case. In the early days of HAR, discretization approaches have been explored - primarily motivated by the desire to minimize computational requirements, but also with a view on applications beyond mere recognition, such as, activity discovery, fingerprinting, or large-scale search. Those traditional discretization approaches, however, suffer from substantial loss in precision and resolution in the resulting representations with detrimental effects on downstream tasks. Times have changed and in this paper we propose a return to discretized representations. We adopt and apply recent advancements in Vector Quantization (VQ) to wearables applications, which enables us to directly learn a mapping between short spans of sensor data and a codebook of vectors, resulting in recognition performance that is generally on par with their contemporary, continuous counterparts - sometimes surpassing them. Therefore, this work presents a proof-of-concept for demonstrating how effective discrete representations can be derived, enabling applications beyond mere activity classification but also opening up the field to advanced tools for the analysis of symbolic sequences, as they are known, for example, from domains such as natural language processing. Based on an extensive experimental evaluation on a suite of wearables-based benchmark HAR tasks, we demonstrate the potential of our learned discretization scheme and discuss how discretized sensor data analysis can lead to substantial changes in HAR.
... As a result, there has been recent interest in classifying batter outcomes in a more accessible manner. Some early approaches used motion estimation or wearable technology for different shots (Khan et al., 2017). Like bowling, however, these approaches have mostly been made obsolete with CNNs, particularly in broadcasted cricket matches (Foysal et al., 2019;Harun-ur-Rashid et al., 2018;Khan et al., 2018;Semwal et al., 2018;Sen et al., 2021). ...
... Human activity recognition (HAR) represents one of the core pillars of mobile, wearable, and pervasive computing, with many practical applications in fields such as health and sleep monitoring [1]- [7], behaviour and skill assessment [8]- [11], or sports coaching [12]- [14], to name but a few. For ...
... traditional machine learning based HAR approaches, feature engineering plays an important role [15]- [18]. With expertdesigned features, classification can be performed using conventional classifiers such as SVM or KNN [1], [12]. However, designing effective features tends to be a somewhat tedious trial-and-error process, and discriminant features may vary from task to task, making system-developing expensive and less sustainable. ...
Human Activity Recognition (HAR) is one of the core research areas in mobile and wearable computing. With the application of deep learning (DL) techniques such as CNN, recognizing periodic or static activities (e.g, walking, lying, cycling, etc.) has become a well studied problem. What remains a major challenge though is the sporadic activity recognition (SAR) problem, where activities of interest tend to be non periodic, and occur less frequently when compared with the often large amount of irrelevant background activities. Recent works suggested that sequential DL models (such as LSTMs) have great potential for modeling nonperiodic behaviours, and in this paper we studied some LSTM training strategies for SAR. Specifically, we proposed two simple yet effective LSTM variants, namely delay model and inverse model, for two SAR scenarios (with and without time critical requirement). For time critical SAR, the delay model can effectively exploit predefined delay intervals (within tolerance) in form of contextual information for improved performance. For regular SAR task, the second proposed, inverse model can learn patterns from the time series in an inverse manner, which can be complementary to the forward model (i.e.,LSTM), and combining both can boost the performance. These two LSTM variants are very practical, and they can be deemed as training strategies without alteration of the LSTM fundamentals. We also studied some additional LSTM training strategies, which can further improve the accuracy. We evaluated our models on two SAR and one non-SAR datasets, and the promising results demonstrated the effectiveness of our approaches in HAR applications.
... To the best of our knowledge, not too many literary works have concentrated on the players' performance assessment. Similar to [13], PerMTL also uses multiple body-worn IMU sensors focused on assessing sports activity. However, it differs in that the intuitive assessment is accomplished through a novel multi-tasking approach that operates end-to-end. ...