Subject Cross Validation in Human Activity Recognition

Dehghani, Akbar; Glatard, Tristan; Shihab, Emad

doi:10.48550/arxiv.1904.02666

Cited by 6 publications

(7 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They evaluate classification outcomes using 10-fold CV and LOUOCV, demonstrating that only the latter method is suitable for classifying unseen user data. Similarly, in [59], the authors explore the impact of the subject CV on the performance of human activity recognition. Their findings indicate that k-fold CV tends to overestimate system performance by approximately 16% when overlapping windows are utilized.…”

Section: Evaluation Methodsmentioning

confidence: 99%

Gender Detection Based on Gait Data: A Deep Learning Approach With Synthetic Data Generation and Continuous Wavelet Transform

Davarci,

Anarim

2023

IEEE Access

View full text Add to dashboard Cite

Smart devices equipped with various sensors enable the acquisition of users' behavioral biometrics. These sensor data capture variations in users' interactions with the devices, which can be analyzed to extract valuable information such as user activity, age group, and gender. In this study, we investigate the feasibility of using gait data for gender detection of users. To achieve this, we propose a novel gender detection scheme based on a deep learning approach, incorporating synthetic data generation and continuous wavelet transform (CWT). In this scheme, the real dataset is first divided into training and test datasets, and then synthetic data are intelligently generated using various techniques to augment the existing training data. Subsequently, CWT is used as the feature extraction module, and its outputs are fed into a deep learning model to detect the gender of users. Different deep learning models, including convolutional neural network (CNN) and long short-term memory (LSTM), are employed in classification. Consequently, we evaluate our proposed framework on different publicly available datasets. On the BOUN Sensor dataset, we obtain an accuracy of 94.83%, marking a substantial 6.5% enhancement over the prior highest rate of 88.33%. Additionally, we achieve 86.27% and 88.15% accuracy on the OU-ISIR Android and OU-ISIR Center IMUZ datasets, respectively. Our experimental results demonstrate that our proposed model achieves high detection rates and outperforms previous methods across all datasets. INDEX TERMSBiometrics, continuous wavelet transform, convolutional neural networks, frequency domain, gender detection, generative adversarial networks, human gait, motion sensors, smartphones. authors successfully determined user gender by analyzing smartphone accelerometer sensor data. Accelerometer sen-sor data have also been leveraged to detect behaviors associated with individuals' stress levels in [2]. Additionally, smartphone motion sensor data has been utilized in multiple studies to recognize and classify human daily activities [3]-[5]. On the other hand, the authors have employed sensor data from smart devices to enhance user identification and authentication systems in [6], [7].Gait analysis is a significant biometric feature that facilitates human identification and provides insights into physical and medical conditions. Due to the unique nature of an individual's gait, which reflects their walking style and physical abilities, it becomes challenging to mimic the gait pattern of others [8]. Thus, gait analysis finds applications across various domains, including security, sports, surveillance, and the medical field [9], [10]. For instance, in [11], the authors underscored using sensor-based gait analysis in clinical applications for monitoring and diagnosing conditions such as Parkinson's disease.Sensor data originating from touchscreen interactions and

show abstract

Section: Evaluation Methodsmentioning

confidence: 99%

Gender Detection Based on Gait Data: A Deep Learning Approach With Synthetic Data Generation and Continuous Wavelet Transform

Davarci,

Anarim

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Taking into account the evaluation value of the prediction result, such as the correct answer rate of the output classification and the regression value, we examine whether the model has actual versatility. Typical cross validation methods include the holdout method and the k-fold method [12,13]. The k-fold method is a cross validation method that divides the training data into several (k) pieces and repeats model construction and verification for the number of divided data (i.e., k times).…”

Section: Evaluation Of Predictive Models and Predictive Performancementioning

confidence: 99%

Ensemble Learning, Deep Learning-Based and Molecular Descriptor-Based Quantitative Structure–Activity Relationships

Matsuzaka

Uesawa

2023

Molecules

View full text Add to dashboard Cite

A deep learning-based quantitative structure–activity relationship analysis, namely the molecular image-based DeepSNAP–deep learning method, can successfully and automatically capture the spatial and temporal features in an image generated from a three-dimensional (3D) structure of a chemical compound. It allows building high-performance prediction models without extracting and selecting features because of its powerful feature discrimination capability. Deep learning (DL) is based on a neural network with multiple intermediate layers that makes it possible to solve highly complex problems and improve the prediction accuracy by increasing the number of hidden layers. However, DL models are too complex when it comes to understanding the derivation of predictions. Instead, molecular descriptor-based machine learning has clear features owing to the selection and analysis of features. However, molecular descriptor-based machine learning has some limitations in terms of prediction performance, calculation cost, feature selection, etc., while the DeepSNAP–deep learning method outperforms molecular descriptor-based machine learning due to the utilization of 3D structure information and the advanced computer processing power of DL.

show abstract

“…Moreover, none of the existing studies reported perfect or near-perfect accuracy, which is important for such sensitive usage. In fact, the adoption of cross-validation alone to measure the performance of the models of almost all the studies [22]- [24], raise serious concern as to whether they truly reflect the models' realistic performances [35]. Additionally, the accuracy of HAR solution for Salat might hamper if a person performs any extra activity in Salat, that does not nullify prayer [36] as shown in Figure 1.…”

Section: Motivations Behind Our Studymentioning

confidence: 99%

Leveraging a Smartwatch for Activity Recognition in Salat

Jahan,

Al-Nabhan,

Noor

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Salat, the most important worship of Muslims and the second pillar of Islam, is an integral part of the Muslim community. Being a complex activity, Salat involves a series of steady and transitional activities to be performed in a specific sequence. On top of that, Salat has variations based on time, priority, school of thought, etc., making activity recognition in Salat more challenging. Existing research studies related to recognizing individual activities in Salat either demand capturing images by a camera or carrying a smartphone (sometimes in inconvenient places) while praying. Both of the demands are not convenient or applicable in real cases. Besides, the existing studies lack user-independent accuracy analysis and fine-grained prediction. To address these gaps, in this study, we first assess the requirement and acceptability of technological solutions for activity recognition in Salat by conducting an exploratory study. Upon establishing the requirement, we propose an activity recognition methodology using a smartwatch to recognize different activities in Salat. We prepare a Salat activity dataset using a smartwatch and propose a new methodology using semantic rules and Dynamic Time Warping (DTW) that achieves a near-perfect accuracy (99.3%) in recognizing activities in Salat. Besides, our proposed methodology offers fine-grained recognition of the individual activities in Salat and is robust enough to overlook the extra transitional activities a person performs while praying, which does not nullify Salat. Therefore, this research is expected to lead to a comprehensive solution for monitoring Salat.

show abstract

Subject Cross Validation in Human Activity Recognition

Cited by 6 publications

References 8 publications

Gender Detection Based on Gait Data: A Deep Learning Approach With Synthetic Data Generation and Continuous Wavelet Transform

Gender Detection Based on Gait Data: A Deep Learning Approach With Synthetic Data Generation and Continuous Wavelet Transform

Ensemble Learning, Deep Learning-Based and Molecular Descriptor-Based Quantitative Structure–Activity Relationships

Leveraging a Smartwatch for Activity Recognition in Salat

Contact Info

Product

Resources

About