Research on 3D virtual vision matching based on interactive color segmentation.

Given the prevalent issues surrounding accuracy and efficiency in contemporary stereo-matching algorithms, this research introduces an innovative image segmentation-based approach. The proposed methodology integrates residual and Swim Transformer modules into the established 3D Unet framework, yielding the Res-Swim-UNet image segmentation model. The algorithm estimates the disparateness of segmented outputs by employing regression techniques, culminating in a comprehensive disparity map.

[Research on portable airway impedance monitoring device based on expiratory oscillation].

Monitoring airway impedance has significant clinical value in accurately assessing and diagnosing pulmonary function diseases at an early stage. To address the issue of large oscillator size and high power consumption in current pulmonary function devices, this study adopts a new strategy of expiration-driven oscillation. A lightweight and low-power airway impedance monitoring system with integrated sensing, control circuitry, and dynamic feedback system, providing visual feedback on the system's status, was developed.