A novel deep learning approach using blurring image techniques for Bluetooth-based indoor localisation

Castillo-Cara, Manuel; Talla-Chumpitaz, Reewos; Orozco-Barbosa, Luis; García‐Castro, Raúl

doi:10.1016/j.inffus.2022.10.011

Cited by 14 publications

(4 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…6 of Supplement File 1 for more details about these techniques). These manifold mappings can be further enhanced by applying blurring techniques 23 , Gabor filtering 24 and the assignment distribution algorithm 53 . The DeepInsight pipeline incorporates several additional steps to enhance the transformation process.…”

Section: Transformation Phase Of Mrep-deepinsightmentioning

confidence: 99%

“…The core motivation behind MRep-DeepInsight is to leverage the complementary nature of different data representations. By simultaneously employing various established feature mapping techniques, such as UMAP 12 , t-SNE 11 , blurring technique 23 , and Gabor filtering 24 , we generate multiple representations of each sample, thereby providing a comprehensive view of its underlying characteristics. Theoretically, with an infinite number of manifold techniques, we could represent any nonlinear function, reiterating the principle in neural networks where an infinite number of nodes in middle layers allows for the representation of any function.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Enhanced analysis of tabular data through Multi-representation DeepInsight

Sharma,

López,

Jia

et al. 2024

Sci Rep

View full text Add to dashboard Cite

Tabular data analysis is a critical task in various domains, enabling us to uncover valuable insights from structured datasets. While traditional machine learning methods can be used for feature engineering and dimensionality reduction, they often struggle to capture the intricate relationships and dependencies within real-world datasets. In this paper, we present Multi-representation DeepInsight (MRep-DeepInsight), a novel extension of the DeepInsight method designed to enhance the analysis of tabular data. By generating multiple representations of samples using diverse feature extraction techniques, our approach is able to capture a broader range of features and reveal deeper insights. We demonstrate the effectiveness of MRep-DeepInsight on single-cell datasets, Alzheimer's data, and artificial data, showcasing an improved accuracy over the original DeepInsight approach and machine learning methods like random forest, XGBoost, LightGBM, FT-Transformer and L2-regularized logistic regression. Our results highlight the value of incorporating multiple representations for robust and accurate tabular data analysis. By leveraging the power of diverse representations, MRep-DeepInsight offers a promising new avenue for advancing decision-making and scientific discovery across a wide range of fields.

show abstract

Section: Transformation Phase Of Mrep-deepinsightmentioning

confidence: 99%

mentioning

confidence: 99%

Enhanced analysis of tabular data through Multi-representation DeepInsight

Sharma,

López,

Jia

et al. 2024

Sci Rep

View full text Add to dashboard Cite

show abstract

“…During the training of the network, feedforward operations are performed layer by layer on the input data, the final result of the output operation is compared with the target, and the error is returned to each neuron to https://doi.org/10.17993/3ctic.2023.122.97-115 adjust the connection parameters of each neuron. Multilayer neural network connections can accurately represent the complex nonlinear mapping relationships between input and output data [19][20][21]. It is similar to the internalization and construction of knowledge in the human learning process.…”

Section: Deep Learningmentioning

confidence: 99%

Research on the Realization Mechanism and Evaluation System of High-Quality Undergraduate Education in Private Universities Based on Deep Learning

Wang

2023

3C TIC

View full text Add to dashboard Cite

Due to the new development stage, it is especially important to improve the education quality of private undergraduate universities. As a result, it is a new hot issue for the construction of a mechanism and assessment system for the quality improvement of private undergraduate education. In this paper, after analyzing and researching the quality of undergraduate education in present-day universities, the mechanism of deep learning is applied to the establishment of the assessment system. Finally, 1082 samples collected from the data center platform of a private university are analyzed as the research object. From the results, the final size of the combined weights of the seven evaluation items constituting the assessment system differed basically little. They were 12.81%, 15.78%, 15.28%, 14.38%, 12.83%, 12.81%, 15.01%, and 13.27%, respectively. In the comparison of this paper's method with FAHP+TOPSIS combined evaluation, euclidean map method, and genetic algorithm assignment, the difference between the seven weight values of the euclidean map method is larger, 5.56%. The evaluation times of the four methods were 41 s, 38 s, 47 s, and 118 s. Compared with the other three methods, the genetic algorithm assignment took the most time.

show abstract

“…Therefore, we propose to incorporate indoor location information into indoor HAR. Compared with indoor positioning technologies such as UWB and Bluetooth, geomagnetic positioning technology can rely solely on the built-in sensors of smartphones to achieve positioning, without the need to deploy additional infrastructure in the environment in advance [ 6 ]. This can not only reduce deployment costs but also increase the practicality of the system.…”

Section: Introductionmentioning

confidence: 99%

Position-Aware Indoor Human Activity Recognition Using Multisensors Embedded in Smartphones

Wang,

2024

Sensors

View full text Add to dashboard Cite

Composite indoor human activity recognition is very important in elderly health monitoring and is more difficult than identifying individual human movements. This article proposes a sensor-based human indoor activity recognition method that integrates indoor positioning. Convolutional neural networks are used to extract spatial information contained in geomagnetic sensors and ambient light sensors, while transform encoders are used to extract temporal motion features collected by gyroscopes and accelerometers. We established an indoor activity recognition model with a multimodal feature fusion structure. In order to explore the possibility of using only smartphones to complete the above tasks, we collected and established a multisensor indoor activity dataset. Extensive experiments verified the effectiveness of the proposed method. Compared with algorithms that do not consider the location information, our method has a 13.65% improvement in recognition accuracy.

show abstract

A novel deep learning approach using blurring image techniques for Bluetooth-based indoor localisation

Cited by 14 publications

References 36 publications

Enhanced analysis of tabular data through Multi-representation DeepInsight

Enhanced analysis of tabular data through Multi-representation DeepInsight

Research on the Realization Mechanism and Evaluation System of High-Quality Undergraduate Education in Private Universities Based on Deep Learning

Position-Aware Indoor Human Activity Recognition Using Multisensors Embedded in Smartphones

Contact Info

Product

Resources

About