The future 6G network will enhance smart cities' data integration and analysis capabilities through Integrated Sensing and Communication (ISAC) technology. This paper proposes a millimeter wave (mmW) in-band full-duplex (IBFD) ISAC system architecture and explores three direction-finding techniques based on subarray splitting for its architecture. Firstly, the mutual interference problem in the IBFD ISAC system was analyzed. Using the Taylor synthesis method to suppress array sidelobes and reduce false alarm rate, combined with a two-dimensional fast Fourier transform (2D-FFT), the direction-finding accuracy is improved under an extremely low signal-to-noise ratio (SNR). Then, the performance of three direction-finding methods was compared through simulation. The simulation results showed that the sum difference beam direction-finding method has better mean square error (MSE) performance. In addition, a sum difference beam direction-finding experiment was conducted, and the direction-finding errors for moving and static targets were as low as 0.6° and 0.5°, respectively, demonstrating excellent direction-finding capability.

Deep learning-based disease grading technologies facilitate timely medical intervention due to their high efficiency and accuracy. Recent advancements have enhanced grading performance by incorporating the ordinal relationships of disease labels. However, existing methods often assume same probability distributions for disease labels across instances within the same category, overlooking variations in label distributions. Additionally, the hyperparameters of these distributions are typically determined empirically, which may not accurately reflect the true distribution. To address these limitations, we propose a disease grading network utilizing a sample-aware asymmetric Gaussian label distribution, termed DGN-AGLD. This approach includes a variance predictor designed to learn and predict parameters that control the asymmetry of the Gaussian distribution, enabling distinct label distributions within the same category. This module can be seamlessly integrated into standard deep learning networks. Experimental results on four disease datasets validate the effectiveness and superiority of the proposed method, particularly on the IDRiD dataset, where it achieves a diabetic retinopathy accuracy of 77.67%. Furthermore, our method extends to joint disease grading tasks, yielding superior results and demonstrating significant generalization capabilities. Visual analysis indicates that our method more accurately captures the trend of disease progression by leveraging the asymmetry in label distribution. Our code is publicly available on https://github.com/ahtwq/AGNet.