Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 2 appraisals
PloS one
Deep learning-based bimodal speech and facial expression recognition of miners' unsafe emotions
Under the influence of unsafe emotions, miners' ability to perceive risks is hindered, which can easily lead to decision-making errors and safety accidents. To recognize unsafe emotions exhibited by miners during operations, this study proposes a deep learning-based bimodal framework that integrates speech and facial expression features. A convolutional neural network (CNN) combined with a bidirectional long short-term memory (Bi-LSTM) network is employed to model local spectral patterns and temporal dependencies in speech signals, and ShuffleNet-V2 is used to capture deep facial features. In addition, three feature enhancement strategies are proposed to improve the generalization ability of the model. By constructing a dataset containing five categories of miners' unsafe emotions for network training, the model achieves a mean recognition accuracy of 85.56%. Furthermore, we conducted a preliminary field test of the bimodal model in a real mining environment. The results provide preliminary evidence of its potential applicability in real-world mining conditions.
16 May 2026
Read appraisal →Scientific reports
Predictive robot eyes shape visual attention, performance, and trust in interaction with an industrial CoBot.
Collaborative robots increasingly share workspaces with humans, making the predictability of robot actions critical for efficient and safe coordination. Predictive visual cues, such as gaze or arrows, may reduce uncertainty, yet their effectiveness and robustness, especially under failure conditions, remain insufficiently understood. Across two laboratory studies with an industrial CoBot, we examined how predictive cues influence visual attention, prediction, and trust. Participants predicted the movement target of a robotic arm displaying eyes, arrows, or no cues across pre-failure, failure, and post-failure phases, while eye movements, reaction times, and trust were recorded. The second study served as a full replication. Across both studies, gaze cues accelerated visual orienting toward the target, as reflected in earlier target fixations, and improved prediction time when cues were reliable. A cue-action mismatch led to a temporary delayed target fixation and slower responses in both cue conditions. Importantly, performance and attentional guidance largely recovered during subsequent error-free interaction, with gaze cues tending to regain their advantage over no cues. Trust mostly showed the expected dynamic pattern of formation, dissolution, and restoration. Together, the findings demonstrate that predictive cues in industrial HRI enhance action anticipation by guiding visual attention, but their benefits are sensitive to cue reliability.
11 May 2026
Read appraisal →