Action Recognition Performance Research Articles

Segment-Based Unsupervised Deep Learning for Human Activity Recognition using Accelerometer Data and SBOA based Channel Attention Networks

The deep learning community has increasingly focused on the critical challenges of human activity segmentation and detection based on sensors, which have numerous real-world applications. In most prior efforts, activity segmentation and recognition have been treated as separate processes, relying on pre-segmented sensor streams. This research proposes an unsupervised deep learning approach for Human Activity Recognition (HAR) that is segment-based, with an emphasis on activity continuity. The approach integrates segment-based SimCLR with Segment Feature Decorrelation (SDFD) and a new framework that leverages pairs of segment data for contrastive learning of visual representations. Furthermore, the Secretary Bird Optimization Algorithm (SBOA) and Channel Attention with Spatial Attention Network (CASANet) are utilized to enhance the performance of sensor-based human activity detection. CASANet effectively extracts key features and spatial dependencies in sensor data, while SBOA optimizes the model for greater accuracy and generalization. Evaluations on two publicly available datasets—Mhealth and PAMAP2—demonstrated an average F1 score of 98%, highlighting the approach’s efficacy in improving activity recognition performance.

Generalized and Improved Human Activity Recognition for Real-Time Wellness Monitoring

Human activity categorization using smartphone data can be useful for physicians in real-time data monitoring in sports or lifestyle monitoring. The goal of this research is to develop a methodology that can identify strong machine-learning classifiers applied to various human activity datasets. The first step is pre-processing the data, followed by feature extraction, selection, and classification. Relying on a single dataset does not yield high confidence in the findings. Instead, examining multiple datasets is crucial for a comprehensive understanding, as it avoids the pitfalls of basing conclusions on one dataset alone. Multiple datasets and classifiers are applied in different experiments to achieve improved and generalized human activity recognition performance. Experimental results of the support vector machine (SVM) with its generalized performance of 99% encourage us to use the trained SVM-based model to monitor normal human activities inside the home, in the park, in the gym, etc. enhancing wellness monitoring.

InfoGCN++: Learning Representation by Predicting the Future for Online Skeleton-based Action Recognition.

Skeleton-based action recognition has made significant advancements recently, with models like InfoGCN showcasing remarkable accuracy. However, these models exhibit a key limitation: they necessitate complete action observation prior to classification, which constrains their applicability in realtime situations such as surveillance and robotic systems. To overcome this barrier, we introduce InfoGCN++, an innovative extension of InfoGCN, explicitly developed for online skeletonbased action recognition. InfoGCN++ augments the abilities of the original InfoGCN model by allowing real-time categorization of action types, independent of the observation sequence's length. It transcends conventional approaches by learning from current and anticipated future movements, thereby creating a more thorough representation of the entire sequence. Our approach to prediction is managed as an extrapolation issue, grounded on observed actions. To enable this, InfoGCN++ incorporates Neural Ordinary Differential Equations, a concept that lets it effectively model the continuous evolution of hidden states. Following rigorous evaluations on three skeleton-based action recognition benchmarks, InfoGCN++ demonstrates exceptional performance in online action recognition. It consistently equals or exceeds existing techniques, highlighting its significant potential to reshape the landscape of real-time action recognition applications. Consequently, this work represents a major leap forward from InfoGCN, pushing the limits of what's possible in online, skeleton-based action recognition.

Hybrid attentive prototypical network for few-shot action recognition

Most previous few-shot action recognition works tend to process video temporal and spatial features separately, resulting in insufficient extraction of comprehensive features. In this paper, a novel hybrid attentive prototypical network (HAPN) framework for few-shot action recognition is proposed. Distinguished by its joint processing of temporal and spatial information, the HAPN framework strategically manipulates these dimensions from feature extraction to the attention module, consequently enhancing its ability to perform action recognition tasks. Our framework utilizes the R(2+1)D backbone network, coupling the extraction of integrated temporal and spatial features to ensure a comprehensive understanding of video content. Additionally, our framework introduces the novel Residual Tri-dimensional Attention (ResTriDA) mechanism, specifically designed to augment feature information across the temporal, spatial, and channel dimensions. ResTriDA dynamically enhances crucial aspects of video features by amplifying significant channel-wise features for action distinction, accentuating spatial details vital for capturing the essence of actions within frames, and emphasizing temporal dynamics to capture movement over time. We further propose a prototypical attentive matching module (PAM) built on the concept of metric learning to resolve the overfitting issue common in few-shot tasks. We evaluate our HAPN framework on three classical few-shot action recognition datasets: Kinetics-100, UCF101, and HMDB51. The results indicate that our framework significantly outperformed state-of-the-art methods. Notably, the 1-shot task, demonstrated an increase of 9.8% in accuracy on UCF101 and improvements of 3.9% on HMDB51 and 12.4% on Kinetics-100. These gains confirm the robustness and effectiveness of our approach in leveraging limited data for precise action recognition.

A Review of Deep Learning Based Video Action Recognition Techniques

Video action recognition is an important research direction in the field of computer vision and pattern recognition, with extensive applications in intelligent video surveillance, human-computer interaction, and sports analysis. The development of data storage and computing hardware over the past decade has driven a shift from traditional feature extraction and machine learning algorithms to deep learning-based approaches. This paper reviews the current state of development, problems, and future research directions of video action recognition techniques. Traditional methods are gradually being replaced by deep learning methods such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), and long-short-term memory networks (LSTMs). These methods automatically extract features and handle time-dependency, significantly improving the accuracy and robustness of action recognition. In particular, models based on the attention mechanism further enhance action recognition performance by dynamically adjusting the focus of attention, a current hot spot in research. Despite many advances, video action recognition still faces several challenges, including high computational resource requirements, complex model training, dataset bias issues, and variations in real-world application scenarios such as viewpoint changes, lighting changes, and occlusion. Future research can explore multi-modal fusion, lightweight models, self-supervised learning, and cross-domain transfer learning to improve the accuracy, robustness, and generalization of action recognition. The review provided aims to offer researchers a comprehensive perspective on the current state of development and future research directions of video action recognition technology.

Attention-guided mask learning for self-supervised 3D action recognition

Most existing 3D action recognition works rely on the supervised learning paradigm, yet the limited availability of annotated data limits the full potential of encoding networks. As a result, effective self-supervised pre-training strategies have been actively researched. In this paper, we target to explore a self-supervised learning approach for 3D action recognition, and propose the Attention-guided Mask Learning (AML) scheme. Specifically, the dropping mechanism is introduced into contrastive learning to develop Attention-guided Mask (AM) module as well as mask learning strategy, respectively. The AM module leverages the spatial and temporal attention to guide the corresponding features masking, so as to produce the masked contrastive object. The mask learning strategy enables the model to discriminate different actions even with important features masked, which makes action representation learning more discriminative. What’s more, to alleviate the strict positive constraint that would hinder representation learning, the positive-enhanced learning strategy is leveraged in the second-stage training. Extensive experiments on NTU-60, NTU-120, and PKU-MMD datasets show that the proposed AML scheme improves the performance in self-supervised 3D action recognition, achieving state-of-the-art results.

A Robust Deep Feature Extraction Method for Human Activity Recognition Using a Wavelet Based Spectral Visualisation Technique.

Human Activity Recognition (HAR), alongside Ambient Assisted Living (AAL), are integral components of smart homes, sports, surveillance, and investigation activities. To recognize daily activities, researchers are focusing on lightweight, cost-effective, wearable sensor-based technologies as traditional vision-based technologies lack elderly privacy, a fundamental right of every human. However, it is challenging to extract potential features from 1D multi-sensor data. Thus, this research focuses on extracting distinguishable patterns and deep features from spectral images by time-frequency-domain analysis of 1D multi-sensor data. Wearable sensor data, particularly accelerator and gyroscope data, act as input signals of different daily activities, and provide potential information using time-frequency analysis. This potential time series information is mapped into spectral images through a process called use of 'scalograms', derived from the continuous wavelet transform. The deep activity features are extracted from the activity image using deep learning models such as CNN, MobileNetV3, ResNet, and GoogleNet and subsequently classified using a conventional classifier. To validate the proposed model, SisFall and PAMAP2 benchmark datasets are used. Based on the experimental results, this proposed model shows the optimal performance for activity recognition obtaining an accuracy of 98.4% for SisFall and 98.1% for PAMAP2, using Morlet as the mother wavelet with ResNet-101 and a softmax classifier, and outperforms state-of-the-art algorithms.

A human mesh-centered approach to action recognition in the operating room

Aim: Video review programs in hospitals play a crucial role in optimizing operating room workflows. In scenarios where split-seconds can change the outcome of a surgery, the potential of such programs to improve safety and efficiency is profound. However, leveraging this potential requires a systematic and automated analysis of human actions. Existing methods predominantly employ manual methods, which are labor-intensive, inconsistent, and difficult to scale. Here, we present an AI-based approach to systematically analyze the behavior and actions of individuals from operating rooms (OR) videos. Methods: We designed a novel framework for human mesh recovery from long-duration surgical videos by integrating existing human detection, tracking, and mesh recovery models. We then trained an action recognition model to predict surgical actions from the predicted temporal mesh sequences. To train and evaluate our approach, we annotated an in-house dataset of 864 five-second clips from simulated surgical videos with their corresponding actions. Results: Our best model achieves an F1 score and the area under the precision-recall curve (AUPRC) of 0.81 and 0.85, respectively, demonstrating that human mesh sequences can be successfully used to recover surgical actions from operating room videos. Model ablation studies suggest that action recognition performance is enhanced by composing human mesh representations with lower arm, pelvic, and cranial joints. Conclusion: Our work presents promising opportunities for OR video review programs to study human behavior in a systematic, scalable manner.

Motion-guided spatiotemporal multitask feature discrimination for self-supervised video representation learning

Powerful self-supervised representation models are able to step out of the traditional supervised paradigm and rely merely on unlabeled data to achieve a deep understanding of visual semantic features. However, previous approaches may suffer from background scene bias, making it difficult to perform a comprehensive exploration of video spatiotemporal structure. To address this challenge, this paper proposes a self-supervised video representation learning framework of motion-guided spatiotemporal multitask feature discrimination (MSMFD). The method mainly utilizes the consistency of motion cues between different views to guide the model for spatial and temporal feature similarity discrimination. Specifically, the model first selects video clips with large motion amplitudes based on the collected optical flow maps. Subsequently, the model introduces an instance discrimination task for overall spatiotemporal structure perception of the video, while a shuffled triplet and an augmented quadruple task are created to further enhance the exploration of intraframe sequence order and local spatial fine-grained. Furthermore, we propose joint motion alignment of spatial, temporal, and spatiotemporal dimensions under different views as a powerful compensation for acquiring motion features. Experimental results demonstrate that our self-supervised method is effective for learning video representations and achieves competitive performance in action recognition and video retrieval tasks compared to other state-of-the-art methods.

Efficient human activity recognition: A deep convolutional transformer-based contrastive self-supervised approach using wearable sensors

Artificial intelligence has advanced the applications of sensor-based human motion capture and recognition technology in various engineering fields, such as human–robot collaboration and health monitoring. Deep learning methods can achieve satisfactory recognition results when provided with sufficient labeled data. However, labeling a large dataset is expensive and time-consuming in practical applications. To address this issue, this paper proposes a deep convolutional transformer-based contrastive self-supervised (DCTCSS) model under the bootstrap your own latent (BYOL) framework. The DCTCSS model aims to achieve reliable activity recognition using only a small amount of labeled data. Firstly, a deep convolutional transformer (DCT) model is proposed as the backbone of DCTCSS model, to learn high-level feature representations from unlabeled data in pre-training period. Subsequently, a simple linear classifier is trained with supervised fine-tuning using a limited amount of labeled data to recognize activities. In addition, this paper experimentally formulates a random data augmentation strategy to increase the diversity of input data. The performance of the DCTCSS model is evaluated and compared with several state-of-the-art algorithms on three datasets widely used in daily life, medical monitoring, and intelligent manufacturing. Experimental results show that the DCTCSS model achieves mean F1 scores of 95.64%, 88.39%, and 98.40% on the UCI-HAR, Skoda, and Mhealth datasets, respectively, using only 10% of the labeled data. These results outperform both supervised and unsupervised baseline models. Consequently, the DCTCSS model demonstrates its effectiveness in reducing the dependence on large amounts of labeled data while still achieving competitive activity recognition performance.

Colorimetric sensing based on silver nanoparticle-functionalized copper porphyrin nanozyme for glyphosate

Glyphosate, a widely used herbicide due to the broad-spectrum and high-efficiency properties, has the potential to accumulate in the human body via agricultural food and water sources throughout the food chain. Its safety for both human health and aquatic ecosystems remains a topic of significant debate. In this study, the presence of glyphosate could be directly detected through a notable colorimetric change in an aqueous solution. For this purpose, a silver nanoparticle-functionalized copper porphyrin metal-organic framework (AgNPs/Cu-TCPP) nanozyme was conveniently prepared for colorimetric sensing of glyphosate. The as-synthesized nanozyme that modified AgNPs with Cu-TCPP metal-organic framework as the precursor, has excellent peroxidase mimetic activity and specific recognition performance of glyphosate. The nanozyme demonstrated the ability to catalyze the oxidation of 3,3,5,5-tetramethylbenzidine (TMB) using H2O2, resulting in a marked change in color. When glyphosate was added, which bonded with Cu2+, led to the occupation of the surface-active sites of the nanozyme, thereby making its catalytic performance ineffective. The concentration of glyphosate exhibited a linear range spanning from 2 to 400 μmol L−1, and the limit of detection was 0.67 μmol L−1 (S/N = 3). The colorimetric sensing approach employing AgNPs/Cu-TCPP nanozyme exhibits remarkable selectivity and reproducibility, enabling the detection of real samples with recovery rates of 98.3%–103.5%, indicating its potential applicability in safeguarding water and food quality.

Simulation of artificial intelligence robots in dance action recognition and interaction process based on machine vision

Artificial intelligence robots based on machine vision have shown great potential in dance action recognition and interaction. In order to perform action recognition, a large amount of dance action video data was collected as the basis for training and testing the model. In motion recognition, various computer vision techniques are used to extract motion features from images or videos to capture key information of dance movements. Machine learning technology was adopted to construct an action classification model. By using the data collected in the early stage and the extracted features, an efficient and accurate classifier has been studied and trained. It can classify the input dance actions into different categories, segment continuous video frames into different action segments, and ensure accurate differentiation of different dance actions during the recognition process, avoiding misjudgment or confusion. By accurately capturing and analyzing the posture information of the human body, we can better understand and simulate dance movements, and enhance the performance and realism of artificial intelligence robots in dance interactions. The experimental results show that the system can accurately recognize various dance movements and achieve good interaction with users.

CANet: Comprehensive Attention Network for video-based action recognition

Attention mechanisms play a crucial role in improving action recognition performance. A video, a type of 3D data, can be effectively explored using attention mechanisms from temporal, spatial, and channel dimensions. However, existing methods based on 2D CNN tend to deal with complex spatiotemporal information from one or two of the dimensions, which eventually hampers their overall performance. In this paper, we propose a novel Comprehensive Attention Network (CANet) to model spatiotemporal information in all three dimensions adaptively. CANet is composed of three core plug-and-play components, namely the Global Guided Short-term Motion Module (GG-SMM), the Second-order Guided Long-term Motion Module (SG-LMM), and the Spatial Motion Adaptive Module (SMAM). Specifically, (1) the GG-SMM module is designed to represent local motion clues in the short-term temporal dimension to improve the classification accuracy of fast-tempo actions. (2) The SG-LMM module is designed to jointly motivate fine-grained motion information in the long-term temporal and channel dimensions, thereby facilitating the discrimination of long-term motions. (3) The SMAM module is used to represent motion-sensitive regions in the spatial dimension by learning the spatial object offsets. Extensive experiments have been conducted on four widely used action recognition benchmarks, namely, Something-Something V1, Kinetics-400, UCF-101, and HMDB-51. Experimental results demonstrate that the proposed CANet achieves excellent performance compared with other state-of-the-art methods.

Modality aware contrastive learning for multimodal human activity recognition

SummaryHuman activity recognition is a well‐established research problem in ubiquitous computing. The increased dependency on various smart devices in our daily lives allows us to investigate the sensor data world produced by multimodal sensors embedded in smart devices. However, the raw sensor data are often unlabeled and annotating this vast amount of data are a costly exercise that can often lead to privacy breaches. Self‐supervised learning‐based approaches are at the forefront of learning semantic representation from unlabeled sensor data, including when applied to human activity recognition tasks. As inferring human activity depends on multimodal sensors, addressing the modality difference and inter‐modality dependencies in a model is an important process. This paper proposes a novel self‐supervised learning approach, modality aware contrastive learning (MACL), for representation learning using multimodal sensor data. The approach uses different sensing modalities to create different views of an input signal. Thus, the model is able to learn the representations by maximizing the similarity among different sensing modalities of the same input signal. Extensive experiments were performed on four publicly available human activity recognition data sets to verify the effectiveness of our proposed MACL method. The experimental evaluation results show that the MACL method attains a comparable performance for human activity recognition to the compared baseline models, directly exceeding the performance of models using standard augmentation transformation strategies.

Mixed Resolution Network with hierarchical motion modeling for efficient action recognition

The dual-stream architecture is frequently employed for learning diverse features from videos. This paper introduces a novel Mixed Resolution Network (MixRes) for processing inputs with hybrid spatiotemporal resolutions, namely high-spatial and low-temporal resolution input, as well as low-spatial and high-temporal resolution input. The utilization of mixed spatiotemporal resolutions not only facilitates the independent emphasis of the two streams on appearance and motion encoding but also diminishes the computational burden. Furthermore, by leveraging the characteristics of neural networks with multiple layers, the temporal stream in the proposed network is divided into different steps to capture short-term and long-term motion information. Finally, we design a Temporal Multiscale Motion Excitation (TMME) module, which enhances the motion-related channels of the video representation by employing multiscale temporal differences. We conduct extensive experiments on multiple action recognition benchmarks, including Something-Something V1 & V2 and Kinetics-400. The outcomes validate that the proposed method achieves superior action recognition performance with low computational cost as compared to the state-of-the-art methods.

A hybrid attention-guided ConvNeXt-GRU network for action recognition

In the digital age, with the continuous emergence of large-scale video data, video understanding has become increasingly important. As a core domain, action recognition has garnered widespread attention. However, video exhibits high-dimensional properties and contains human action information at multiple scales, which makes conventional attention mechanisms difficult to capture complex action information. To improve the performance of action recognition, a Hybrid Attention-guided ConvNeXt-GRU Network (HACG) is proposed. Specifically, a Novel Attention Mechanism (ANM) is constructed by integrating a parameter-free attention module into ConvNeXt, enabling the preliminary extraction of important features without the addition of extra parameters. Then, a Multiscale Hybrid Attention Module (MHAM) adopts an improved and efficient Selective Kernel Network (SKNet) to adaptively calibrate channel features. In this way, the module enhances the model’s ability to perceive features at different scales while improving the correlation between channels. Furthermore, MHAM incorporates an Atrous Spatial Pyramid Pooling (ASPP) to extract local and global information from different regions. Finally, MHAM is integrated with the Gated Recurrent Unit (GRU) to capture the interdependence between space and time. Experimental results show that HACG exhibits superior competitiveness compared with the state-of-the-art on the UCF-101, HMDB-51, and Kinetics-400 datasets. This indicates that HACG can more effectively capture important features to suppress noise interference while also having a lower computational load, which makes HACG a highly promising choice for action recognition tasks.

Dynamic Semantic-Based Spatial Graph Convolution Network for Skeleton-Based Human Action Recognition

Graph convolutional networks (GCNs) have attracted great attention and achieved remarkable performance in skeleton-based action recognition. However, most of the previous works are designed to refine skeleton topology without considering the types of different joints and edges, making them infeasible to represent the semantic information. In this paper, we proposed a dynamic semantic-based graph convolution network (DS-GCN) for skeleton-based human action recognition, where the joints and edge types were encoded in the skeleton topology in an implicit way. Specifically, two semantic modules, the joints type-aware adaptive topology and the edge type-aware adaptive topology, were proposed. Combining proposed semantics modules with temporal convolution, a powerful framework named DS-GCN was developed for skeleton-based action recognition. Extensive experiments in two datasets, NTU-RGB+D and Kinetics-400 show that the proposed semantic modules were generalized enough to be utilized in various backbones for boosting recognition accuracy. Meanwhile, the proposed DS-GCN notably outperformed state-of-the-art methods. The code is released here https://github.com/davelailai/DS-GCN

Various frameworks for integrating image and video streams for spatiotemporal information learning employing 2D–3D residual networks for human action recognition

Human action recognition has been identified as an important research topic in computer vision because it is an essential form of communication and interplay between computers and humans to assist computers in automatically recognizing human behaviors and accurately comprehending human intentions. Inspired by some keyframe extraction and multifeatured fusion research, this paper improved the accuracy of action recognition by utilizing keyframe features and fusing them with video features. In this article, we suggest a novel multi-stream approach architecture made up of two distinct models fused using different fusion techniques. The first model combines convolutional neural networks in two-dimensional (2D-CNN) with long-short term memory networks to glean long-term spatial and temporal features from video keyframe images for human action recognition. The second model is a three-dimensional convolutional neural network (3D-CNN) that gathers quick spatial–temporal features from video clips. Subsequently, two frameworks are put forth to explain how various fusion structures can improve the performance of action recognition. We investigate methods for video action recognition using early and late fusion. While the late-fusion framework addresses the decision fusion from the two models' choices for action recognition, the early-fusion framework examines the impact of early feature fusion of the two models for action recognition. The various fusion techniques investigate how much each spatial and temporal feature influences the recognition model's accuracy. The HMDB-51 and UCF-101 datasets are two important action recognition benchmarks used to evaluate our method. When applied to the HMDB-51 dataset and the UCF-101 dataset, the early-fusion strategy achieves an accuracy of 70.1 and 95.5%, respectively, while the late-fusion strategy achieves an accuracy of 77.7 and 97.5%, respectively.

Robust human locomotion and localization activity recognition over multisensory.

Human activity recognition (HAR) plays a pivotal role in various domains, including healthcare, sports, robotics, and security. With the growing popularity of wearable devices, particularly Inertial Measurement Units (IMUs) and Ambient sensors, researchers and engineers have sought to take advantage of these advances to accurately and efficiently detect and classify human activities. This research paper presents an advanced methodology for human activity and localization recognition, utilizing smartphone IMU, Ambient, GPS, and Audio sensor data from two public benchmark datasets: the Opportunity dataset and the Extrasensory dataset. The Opportunity dataset was collected from 12 subjects participating in a range of daily activities, and it captures data from various body-worn and object-associated sensors. The Extrasensory dataset features data from 60 participants, including thousands of data samples from smartphone and smartwatch sensors, labeled with a wide array of human activities. Our study incorporates novel feature extraction techniques for signal, GPS, and audio sensor data. Specifically, for localization, GPS, audio, and IMU sensors are utilized, while IMU and Ambient sensors are employed for locomotion activity recognition. To achieve accurate activity classification, state-of-the-art deep learning techniques, such as convolutional neural networks (CNN) and long short-term memory (LSTM), have been explored. For indoor/outdoor activities, CNNs are applied, while LSTMs are utilized for locomotion activity recognition. The proposed system has been evaluated using the k-fold cross-validation method, achieving accuracy rates of 97% and 89% for locomotion activity over the Opportunity and Extrasensory datasets, respectively, and 96% for indoor/outdoor activity over the Extrasensory dataset. These results highlight the efficiency of our methodology in accurately detecting various human activities, showing its potential for real-world applications. Moreover, the research paper introduces a hybrid system that combines machine learning and deep learning features, enhancing activity recognition performance by leveraging the strengths of both approaches.

Deep neural networks for wearable sensor-based activity recognition in Parkinson’s disease: investigating generalizability and model complexity

BackgroundThe research gap addressed in this study is the applicability of deep neural network (NN) models on wearable sensor data to recognize different activities performed by patients with Parkinson’s Disease (PwPD) and the generalizability of these models to PwPD using labeled healthy data.MethodsThe experiments were carried out utilizing three datasets containing wearable motion sensor readings on common activities of daily living. The collected readings were from two accelerometer sensors. PAMAP2 and MHEALTH are publicly available datasets collected from 10 and 9 healthy, young subjects, respectively. A private dataset of a similar nature collected from 14 PwPD patients was utilized as well. Deep NN models were implemented with varying levels of complexity to investigate the impact of data augmentation, manual axis reorientation, model complexity, and domain adaptation on activity recognition performance.ResultsA moderately complex model trained on the augmented PAMAP2 dataset and adapted to the Parkinson domain using domain adaptation achieved the best activity recognition performance with an accuracy of 73.02%, which was significantly higher than the accuracy of 63% reported in previous studies. The model’s F1 score of 49.79% significantly improved compared to the best cross-testing of 33.66% F1 score with only data augmentation and 2.88% F1 score without data augmentation or domain adaptation.ConclusionThese findings suggest that deep NN models originating on healthy data have the potential to recognize activities performed by PwPD accurately and that data augmentation and domain adaptation can improve the generalizability of models in the healthy-to-PwPD transfer scenario. The simple/moderately complex architectures tested in this study could generalize better to the PwPD domain when trained on a healthy dataset compared to the most complex architectures used. The findings of this study could contribute to the development of accurate wearable-based activity monitoring solutions for PwPD, improving clinical decision-making and patient outcomes based on patient activity levels.