This study aims to capture subtle changes in the pupil, identify relatively weak inter-class changes, extract more abstract and discriminative pupil features, and study a pupil refinement recognition method based on attention mechanisms. Based on the deep learning framework and the ResNet101 deep residual network as the backbone network, a pupil refinement recognition model is established. Among them, the image preprocessing module is used to preprocess the pupil images captured by infrared spectroscopy, removing internal noise from the pupil images. By using the ResNet101 backbone network, subtle changes in the pupil are captured, weak inter-class changes are identified, and different features of the pupil image are extracted. The channel attention module is used to screen pupil features and obtain key pupil features. External attention modules are used to enhance the expression of key pupil feature information and extract more abstract and discriminative pupil features. The Softmax classifier is used to process the pupil features captured by infrared spectra and output refined pupil recognition results. Experimental results show that this method can effectively preprocess pupil images captured by infrared spectroscopy and extract pupil features. This method can effectively achieve fine pupil recognition, and the fine recognition effect is relatively good.

To address the issues of low perception accuracy and poor robustness in traditional motion recognition methods within complex walking environments for visually impaired individuals, this study utilizes multi-modal data, including ECG, PPG, and IMU, for classification. Regarding the low filtering efficiency of multi-modal data, an improved wavelet filtering algorithm based on LSTM is proposed. To further enhance classification accuracy, this paper introduces a motion recognition method for the blindfolded mobility simulation based on an Attention-based Two-Stream Deep Fusion Convolutional Neural Network (ATS-DFCNN). The proposed method constructs a two-stream heterogeneous feature extraction architecture by synchronously collecting tri-axial motion signals and physiological signals from subjects. A 1D-CNN is employed to capture the spatial geometric features of limb movements, while a hybrid CNN-GRU network is utilized to mine the temporal evolution patterns of physiological stress. Furthermore, an attention mechanism is introduced to achieve dynamic weighted fusion at the feature level, which strengthens critical motion features and suppresses environmental noise. Experiments were conducted with 10 subjects simulating the movements of visually impaired individuals, covering typical actions such as walking, standing, climbing stairs, descending stairs, and falling. The results demonstrate that the proposed adaptive filtering algorithm achieves an AUC of 0.942, significantly improving feature distinctiveness compared to traditional algorithms. The ATS-DFCNN model achieved an average recognition accuracy of 92.2% across five activity categories, representing a 4.8% performance increase over single IMU modal classification. Particularly in fall detection, the model effectively reduces false alarms through physiological feedback and accurately infers motion intentions, providing reliable technical support for the safety monitoring of intelligent walking-aid systems.

In aviation safety, pilots must efficiently process dynamic visual information and maintain a high level of attention. Any missed judgment of critical information or delay in decision-making may lead to mission failure or catastrophic consequences. Therefore, accurately detecting pilots’ attention states is the primary prerequisite for improving flight safety and performance. To better detect the attention state of pilots, this paper takes flight trainees as the research object and the simulated flight environment as the experimental background. It proposes a method for detecting the attention state of flight trainees based on a multi-modal dynamic depth network (M3D-Net). The M3D-Net architecture is a lightweight neural network architecture that integrates temporal image features, visual information features, and flight operation data features. It aligns image and text features through an attention mechanism to enhance the semantic association between modalities; it utilizes the Depth-wise Separable Convolution and LSTM (DSC-LSTM) module to model temporal information, dynamically capturing the contextual dependencies within the sequence, and achieving six-level attention state classification. This paper conducted ablation experiments to comparatively analyze the classification effects of the model and also evaluates the effectiveness of our proposed method through model evaluation metrics. Experiments show that the classification effect of the model architecture proposed in this paper reaches 97.56%, with a model size of 18.6 M. Compared with traditional algorithms, the M3D-Net architecture has better performance prospects in terms of application.