The emergence of low-cost edge devices has enabled the integration of automatic speech recognition (ASR) into IoT environments, creating new opportunities for real-time language assessment. However, achieving reliable performance on resource-constrained hardware remains a significant challenge, especially on the Artificial Internet of Things (AIoT). This study presents an AIoT-based framework for automated English-speaking assessment that integrates architecture and system design, ASR benchmarking, and reliability analysis on edge devices. The proposed AIoT-oriented architecture incorporates a lightweight scoring framework capable of analyzing pronunciation, fluency, prosody, and CEFR-aligned speaking proficiency within an automated assessment system. Seven open-source ASR models—four Whisper variants (tiny, base, small, and medium) and three Vosk models—were systematically benchmarked in terms of recognition accuracy, inference latency, and computational efficiency. Experimental results indicate that Whisper-medium deployed on the Raspberry Pi 5 achieved the strongest overall performance, reducing inference latency by 42–48% compared with the Raspberry Pi 4 and attaining the lowest Word Error Rate (WER) of 6.8%. In contrast, smaller models such as Whisper-tiny, with a WER of 26.7%, exhibited two- to threefold higher scoring variability, demonstrating how recognition errors propagate into automated assessment reliability. System-level testing revealed that the Raspberry Pi 5 can sustain near real-time processing with approximately 58% CPU utilization and around 1.2 GB of memory, whereas the Raspberry Pi 4 frequently approaches practical operational limits under comparable workloads. Validation using real learner speech data (approximately 100 sessions) confirmed that the proposed system delivers accurate, portable, and privacy-preserving speaking assessment using low-power edge hardware. Overall, this work introduces a practical AIoT-based assessment framework, provides a comprehensive benchmark of open-source ASR models on edge platforms, and offers empirical insights into the trade-offs among recognition accuracy, inference latency, and scoring stability in edge-based ASR deployments.

Thermal hazards in industrial environments often remain undetected until critical failure or injury occurs. Conventional handheld infrared cameras require manual operation and limit continuous situational awareness. This study presents the design and field validation of a wearable helmet-mounted real-time thermal system based on the MLX90640 infrared array sensor and a Raspberry Pi 5 edge computing platform. Experimental validation was performed across multiple scenarios of 400 measurements based on industrial distances of 100 cm and 150 cm. The performance of the system was tested against a pre-calibrated hotspot infrared thermometer using linear regression analysis and standard error metrics to determine proportional agreement. The results indicate a strong proportional relationship between the two systems at both industrial distances, with R² values ranging from 0.9885 to 0.9973 at 100 cm and from 0.9586 to 0.9867 at 150 cm. A moderate increase in mean absolute error (MAE) was observed as the measurement distance increased. Statistically significant increases in error were identified in mechanically dynamic scenarios where statistically significant increases in measurement error were observed (p-value < 0.05), indicating distance-dependent sensitivity under moving mechanical conditions. The higher absolute errors at longer distances mainly result from field-of-view expansion, reduced target occupancy, and mixed-pixel hotspot effects rather than weakened proportional trend stability. An industrial distance-aware linear regression model was developed to describe behavior and support calibrations under different deployment conditions. Despite minor absolute deviations during dynamic operations, the system maintained strong trend-tracking performance, suggesting suitability for daily preliminary hazard monitoring in industrial safety maintenance.

This study explores the impact and underlying mechanism of Environmental, Social, and Governance (ESG) practices on the financial performance of Chinese listed companies, with a specific focus on the mediating role of corporate innovation. Drawing on stakeholder theory, innovation theory, and the concept of shared value, the research employs panel data from 2011 to 2020 covering 1,380 A-share firms. A series of fixed-effects regressions, mediation models, and instrumental variable tests are conducted to ensure robustness and address endogeneity concerns. The results reveal that ESG governance positively influences firm financial performance, but primarily through its effect on innovation rather than through direct channels. Firms with stronger ESG performance demonstrate higher R\&D intensity and investment in innovation-related personnel, which in turn lead to improved financial outcomes. Mediation analysis confirms that corporate innovation acts as a significant conduit through which ESG practices translate into long-term firm value. Additional findings highlight the role of transparency and risk management as amplifying factors, while excessive agency costs and leverage appear to diminish performance gains. This study contributes to ESG literature by clarifying the internal pathways connecting non-financial governance practices with financial metrics and by contextualizing the ESG–innovation–performance nexus in an emerging market setting. The findings offer strategic implications for corporate managers, investors, and policymakers, emphasizing ESG as a source of innovation-driven competitive advantage. Overall, the research provides a theoretical and empirical foundation for advancing ESG-integrated innovation strategies that foster sustainable financial success.

High-fidelity 3D face reconstruction from a single image is challenging, owing to the inherently ambiguous depth cues and the strong entanglement of multi-scale facial textures. In this regard, we propose a hierarchical multi-resolution self-supervised framework (HMR-Framework), which reconstructs coarse-, medium-, and fine-scale facial geometry progressively through a unified pipeline. A coarse geometric prior is first estimated via 3D morphable model regression, followed by medium-scale refinement using a vertex deformation map constrained by a global–local Markov random field loss to preserve structural coherence. In order to improve fine-scale fidelity, a learnable Gabor-aware texture enhancement module has been proposed to decouple spatial–frequency information and thus improve sensitivity for high-frequency facial attributes. Additionally, we employ a wavelet-based detail perception loss to preserve the edge-aware texture features while mitigating noise commonly observed in in-the-wild images. Extensive qualitative and quantitative evaluation of benchmark datasets indicate that the proposed framework provides better fine-detail reconstruction than existing state-of-the-art methods, while maintaining robustness over pose variations. Notably, the hierarchical design increases semantic consistency across multiple geometric scales, providing a functional solution for high-fidelity 3D face reconstruction from monocular images.

Time-series water-quality forecasting plays a crucial role in sustainable environmental monitoring, early-warning surveillance, and data-driven water-resource governance. Degradation of river ecosystems poses significant risks to public health, biodiversity, and long-term socio-economic resilience, particularly in rapidly developing regions. In this study, a multi-scale diffusion forecaster (MDF) is introduced to enhance predictive accuracy and uncertainty quantification for river water-quality dynamics in Thailand. The proposed framework integrates seasonal-trend decomposition with a hierarchical denoising diffusion process to model stochastic environmental fluctuations across multiple temporal resolutions. Experiments conducted using real water-quality datasets from the Mae Klong, Khwae Noi, and Khwae Yai Rivers, and the Port Authority of Thailand, demonstrate that MDF achieves superior probabilistic calibration under noise and data incompleteness compared to conventional deterministic baselines. The forecasting capability supports proactive pollution control, sustainable resource allocation, and climate-resilient water-policy design, directly contributing to Sustainable Development Goals (SDG 6: Clean Water and Sanitation; SDG 13: Climate Action; and SDG 14: Life Below Water). The findings highlight the potential of diffusion-based learning as an enabling technology for sustainable aquatic ecosystem governance and long-term environmental planning.

This paper introduces an advanced smart sea pontoon designed to enhance sea current behavior monitoring in Thailand's coastal areas. Traditional pontoon systems used in surveying practices face operational challenges and limited durability. In response, the newly developed pontoon integrates Internet of Things (IoT) technology for precise real-time data transmission over long distances, reducing operational duration and minimizing risks to surveying engineer. Constructed with high-density polyethylene or HDPE and reinforced with steel, the pontoon offers superior strength and resistance to withstand marine extreme conditions. Specialized equipment enables convenient data transfer and pontoon monitoring, eliminating the need for physical presence on the pontoon and generating cost savings while providing detailed insights into current speed, direction, temperature, and salinity. This paper explores the design, production, and deployment of the smart sea pontoon, demonstrating its performance in Thailand's marine environments and showcasing its potential to enhance maritime safety and operational efficiency with less cost.