Recent research demonstrates that the fusion of multimodal images can improve the performance of pedestrian detectors under low-illumination environments. However, existing multimodal pedestrian detectors cannot adapt to the variability of environmental illumination. When the lighting conditions of the application environment do not match the experimental data illumination conditions, the detection performance is likely to be stuck significantly. To resolve this problem, we propose a novel transformer–fusion-based YOLO detector to detect pedestrians under various illumination environments, such as nighttime, smog, and heavy rain. Specifically, we develop a novel transformer–fusion module embedded in a two-stream backbone network to robustly integrate the latent interactions between multimodal images (visible and infrared images). This enables the multimodal pedestrian detector to adapt to changing illumination conditions. Experimental results on two well-known datasets demonstrate that the proposed approach exhibits superior performance. The proposed TF-YOLO drastically improves the average precision of the state-of-the-art approach by 3.3% and reduces the miss rate of the state-of-the-art approach by about 6% on the challenging multi-scenario multi-modality dataset.

Myocardial infarction (MI) is a critical cardiovascular disease characterized by extensive myocardial necrosis occurring within a short timeframe. Traditional MI detection and localization techniques predominantly utilize single-domain features as input. However, relying solely on single-domain features of the electrocardiogram (ECG) proves challenging for accurate MI detection and localization due to the inability of these features to fully capture the complexity and variability in cardiac electrical activity. To address this, we propose a multi-domain feature fusion convolutional neural network (MFF–CNN) that integrates the time domain, frequency domain, and time-frequency domain features of ECG for automatic MI detection and localization. Initially, we generate 2D frequency domain and time-frequency domain images to combine with single-dimensional time domain features, forming multi-domain input features to overcome the limitations inherent in single-domain approaches. Subsequently, we introduce a novel MFF–CNN comprising a 1D CNN and two 2D CNNs for multi-domain feature learning and MI detection and localization. The experimental results demonstrate that in rigorous inter-patient validation, our method achieves 99.98% detection accuracy and 84.86% localization accuracy. This represents a 3.43% absolute improvement in detection and a 16.97% enhancement in localization over state-of-the-art methods. We believe that our approach will greatly benefit future research on cardiovascular disease.

Accurate diagnosis of myocardial infarction (MI) holds significant clinical importance for public health systems. Deep learning-based ECG, classification and localization methods can automatically extract features, thereby overcoming the dependence on manual feature extraction in traditional methods. However, these methods still face challenges such as insufficient utilization of dynamic information in cardiac cycles, inadequate ability to capture both global and local features, and data imbalance. To address these issues, this paper proposes a ResNet-Transformer cascaded network (RTCN) to process time frequency features of ECG signals generated by the S-transform. First, the S-transform is applied to adaptively extract global time frequency features from the time frequency domain of ECG signals. Its scalable Gaussian window and high phase resolution can effectively capture the dynamic changes in cardiac cycles that traditional methods often fail to extract. Then, we develop an architecture that combines the Transformer attention mechanism with ResNet to extract multi-scale local features and global temporal dependencies collaboratively. This compensates for the existing deep learning models’ insufficient ability to capture both global and local features simultaneously. To address the data imbalance problem, the Denoising Diffusion Probabilistic Model (DDPM) is applied to synthesize high-quality ECG samples for minority classes, increasing the inter-patient accuracy from 61.66% to 68.39%. Gradient-weighted Class Activation Mapping (Grad-CAM) visualization confirms that the model’s attention areas are highly consistent with pathological features, verifying its clinical interpretability.