Owing to its resilience to visual noise and viewpoint variations, skeleton-based analysis has become a cornerstone of human action recognition research. Despite its practical significance, existing methodologies often suffer from a reliance on single-stream skeletal representations, which fail to encompass the full complexity of action features. This study introduces Latent Features for Human Action Recognition (LFHAR), a novel architecture designed to overcome these limitations by utilizing diverse spatio-temporal latent representations for improved feature extraction. The approach applies graph-based transformations to individual skeletal frames in temporal sequences, then arranges the derived graph features into spatio-temporal matrices. Evaluation of standard datasets demonstrates the stability and invariance characteristics of the LFHAR architecture. The method produces substantial performance improvements, achieving accuracy increases of 2.7% on dataset NTU-RGB+D-60-classes and 2.1% on dataset NTU-RGB+D-120-classes, confirming its accuracy in improving human action recognition.
| Mualliflar | Avazjon Marakhimov, Kabul Khudaybergenov, Mominov Zakhriddin |
|---|---|
| Jurnal | Кимёвий технология. Назорат ва бошқарув |
| Nashr sanasi | 2025-12-28 |
| Jild | 2025 |
| Son | 6 |
| Betlar | 103-109 |
| Til | en |
| DOI | 10.59048/2181-1105.1725 |
DOI: 10.59048/2181-1105.1725 · Maqolaning asl sahifasi
This paper presents a Fuzzy-SPT (Fuzzy Set-Point Tracking) controller designed to combine the strengths of FLC (Fuzzy Logic Control) and PID (Proportional-Integral-Derivative) control while mitigating their limitations…
The article examines methods for improving self-diagnostics of consumption measurement systems based on artificial intelligence in the context of industry digitalization and the development of cyber-physical systems. It…
The article examines the problems and prospects of creating discrete fiber-optic liquid level gauges. Functional diagrams of single-channel and multi-channel discrete fiber-optic liquid level meters are presented…
This article is dedicated to the analogy systems of physical quantities and their graph models in determining the parameters of mechatronic modules. It presents the general sets and fundamental laws of analogy systems…
The article examines modern approaches to modernizing vegetable oil refining processes from the perspective of increasing energy efficiency, environmental friendliness, and final product quality. Traditional and…
In modern digital systems, efficient and reliable information exchange is essential for the stability of corporate systems. Traditional data management models struggle to detect and eliminate invalid, incomplete data at…
This article provides a comparative analysis of the main methods of vegetable oil refining: alkaline (chemical), physical, and combined. It examines the stages of the technological process and the applicability of each…
The article proposes a new analytical method for investigating nonlinear magnetic circuits with a special structure of parameter distribution. The method is based on introducing into the system of nonlinear differential…
This paper considers an approach to adaptive-robust control of dynamic systems using generalized anticipation, where the object is described by a locally linearized model. Known self-tuning algorithms show insufficient…
This article examines the problem of detecting and predicting industrial equipment faults using IoT sensor data through machine learning techniques. Sensor readings such as temperature, vibration, pressure, voltage, and…