Improving Human Activity Recognition Integrating LSTM With Different Data Sources: Features, Object Detection and Skeleton Tracking

Domingo, Jaime Duque; Gómez-García-Bermejo, Jaime; Zalama, Eduardo

doi:10.1109/access.2022.3186465

Cited by 12 publications

(3 citation statements)

References 79 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…shown to assist consumers make better selections, save time searching, and locate the best rates [13]. Software designed to produce and deliver ideas for products or content that a specific user might like to purchase or interact with is called a product recommendation system [14]. The implementation of an e-commerce application for the selling of many sorts of products with similar qualities is the subject of this study.…”

Section: Recommendations Engines Have Beenmentioning

confidence: 99%

Untitled

2023

IJIEMR

View full text Add to dashboard Cite

E-commerce has gradually become one of the most important driving forces for the sustained and vigorous development of country's consumer market. As a result of the rapid development of network and informatization, the application and maturity of technologies like the Internet, terminal equipment, logistics, and payment, and the continuous improvement of people's consumption concepts is growing. The typical marketing methodology does not allow organizations to completely comprehend the needs of their customers. Selling products to customers is the only concern of the e-commerce platforms.They can only concentrate on the product because they don't know what the customer wants.As a result of these limitations, a model for precision marketing that employs integrated convolution neural networks with Bi LSTM (Bidirectional Long Short Term Memory) to create detailed portraits of the target customers. Users' customized items and services are becoming increasingly more dependent on personalized suggestion services as the ecommerce business expands globally. They have the potential to lower the price users pay to gather information while also increasing sales for the company. The use of user reviews to address classic product recommender system research issues has become increasingly common in recent years. Using a convolution and bidirectional long short-term memory hybrid neural network, the review semantics extractor learns how to represent product reviews. Product Recommendations Generator models a user's preference for items based on their prior interactions. In this research a Predictive Integrated Convolution Neural Network with Bi LSTM (PICNN-BiLSTM) is proposed for predictive marketing of e-commerce applications. The proposed model considers Amazon dataset for market prediction and the results represent that the proposed model performance is higher.

show abstract

Section: Recommendations Engines Have Beenmentioning

confidence: 99%

Untitled

2023

IJIEMR

View full text Add to dashboard Cite

show abstract

“…LSTM layer-based neural network models are a type of recurrent network with a number of modifications to prevent the gradient vanishing that conventional RNNs suffer from [ 59 ]. This type of network is also suitable for working with time series and large volumes of data [ 60 ].…”

Section: Activity Recognition Through Deep Learningmentioning

confidence: 99%

SDHAR-HOME: A Sensor Dataset for Human Activity Recognition at Home

Ramos

Domingo

Zalama

et al. 2022

Sensors

Self Cite

View full text Add to dashboard Cite

Nowadays, one of the most important objectives in health research is the improvement of the living conditions and well-being of the elderly, especially those who live alone. These people may experience undesired or dangerous situations in their daily life at home due to physical, sensorial or cognitive limitations, such as forgetting their medication or wrong eating habits. This work focuses on the development of a database in a home, through non-intrusive technology, where several users are residing by combining: a set of non-intrusive sensors which captures events that occur in the house, a positioning system through triangulation using beacons and a system for monitoring the user’s state through activity wristbands. Two months of uninterrupted measurements were obtained on the daily habits of 2 people who live with a pet and receive sporadic visits, in which 18 different types of activities were labelled. In order to validate the data, a system for the real-time recognition of the activities carried out by these residents was developed using different current Deep Learning (DL) techniques based on neural networks, such as Recurrent Neural Networks (RNN), Long Short-Term Memory networks (LSTM) or Gated Recurrent Unit networks (GRU). A personalised prediction model was developed for each user, resulting in hit rates ranging from 88.29% to 90.91%. Finally, a data sharing algorithm has been developed to improve the generalisability of the model and to avoid overtraining the neural network.

show abstract

“…There are two main approaches: data-driven and knowledge-driven. In the data-driven approaches, activities and intentions are probabilistic and statistically represented by using machine learning and data mining techniques (Domingo et al, 2022;Gu et al, 2021;Gomez Cubero and Rehm, 2021;Li and Zhang, 2014). They have shown effective performance in domains with large training sets and demonstrate flexibility for new patterns and robustness to noisy, uncertain, and incomplete data.…”

Section: Introductionmentioning

confidence: 99%

Ontology-based Activity Intention Recognition via Human-in-the-loop Framework

Heo,

Hwang,

Gong

et al. 2024

Preprint

View full text Add to dashboard Cite

Human intention recognition is crucial for various human-centered settings.Existing approaches can be categorized into data-driven and knowledge-drivenmethods. While data-driven methods have shown promising performance withlarge training sets, they face challenges in data collection and limited interpretability.In contrast, knowledge-driven methods have the advantage of notrequiring training data and being interpretable, but acquiring high-qualitydomain knowledge is time-consuming and costly. To address these challenges,we propose OntoIR, a novel Ontology-based human activity Intention Recognitionsystem with a human-in-the-loop rule generation and refinement framework.OntoIR involves an ontology and human-guided logical rules to recognize humanactivities and intentions. By Leveraging rule mining model, the OntoIR extractscandidate rules enriches them with contextual information. Human experts thencurate meaningful rules and supplement insufficient context information. Thisallows us to obtain explainable human intention recognition results throughexpressive rules in situations where data is scarce. To evaluate our system,we collect an Activity/Intention recognition dataset, AICar, in a car showroomenvironment. Comprehensive experiments are conducted on the AICar,demonstrating that the collaboration of rule-mining models and domain expertsimproves the performance of activity intention recognition without lowering theinterpretability of recognition results.

show abstract

Improving Human Activity Recognition Integrating LSTM With Different Data Sources: Features, Object Detection and Skeleton Tracking

Cited by 12 publications

References 79 publications

Untitled

Untitled

SDHAR-HOME: A Sensor Dataset for Human Activity Recognition at Home

Ontology-based Activity Intention Recognition via Human-in-the-loop Framework

Contact Info

Product

Resources

About