4 results about "Digestive tract disease" patented technology

This invention discloses a deformable capsule endoscope probe and image fusion system based on multimodal imaging, belonging to the field of medical endoscopy technology. This invention solves the problems of existing capsule endoscope technology, which has relatively limited functionality and a fixed shape, making it difficult to adapt to the complex structure of the digestive tract and resulting in poor imaging effects. By using flexible silicone material to make the capsule endoscope probe's shell and deformable structure, it can automatically adjust its shape according to the shape of the digestive tract lumen, ensuring clear and accurate multimodal images are acquired from different locations. A neural network model accurately identifies potential lesion areas and adaptively adjusts the fusion weights based on the importance of the lesion and tissue characteristics to optimize image quality. Furthermore, a deep learning fusion network generates a comprehensive image, which, after contrast optimization and automatic labeling of lesion areas, provides doctors with clearer and more detailed diagnostic prompts, aiding in the early diagnosis and precise treatment of digestive tract diseases.

The application provides a gold-silver alloy nanoflower SERS substrate and a preparation method and application thereof. 3+ and Ag + The preparation method is as follows: an aluminum foil paper substrate is placed in a growth solution containing Au 3+ and Ag + , a gold-silver alloy nanoflower array pattern layer with a multi-level branch structure is formed on the surface of the aluminum foil paper substrate by a displacement reaction of aluminum and noble metal ions in situ. The gold-silver alloy nanoflower array pattern layer has rich nanotips and a three-dimensional rough structure, can form a large number of electromagnetic 'hot spots', and can significantly enhance the Raman signal intensity. Relying on the excellent SERS enhancement performance, the gold-silver alloy nanoflower SERS substrate is applied to an immune recognition mediated sandwich detection mode, and high-sensitivity and accurate detection of a hidden blood marker in a fecal sample is realized. The preparation process is simple, the cost is low, and complex equipment is not needed; the obtained SERS substrate has strong stability and excellent repeatability, and can meet the needs of early screening and on-site rapid detection of digestive diseases.

This invention provides a gold-silver alloy nanoflower SERS substrate, its preparation method, and its application. The preparation method involves placing an aluminum foil paper substrate in an Au-containing... 3+ and Ag + In a growth solution, a gold-silver alloy nanoflower array pattern layer with a multi-level branched structure is grown in situ on the surface of an aluminum foil substrate using a displacement reaction between aluminum and noble metal ions. This gold-silver alloy nanoflower array pattern layer has abundant nano-tip and three-dimensional rough structure, which can form a large number of electromagnetic "hot spots" and significantly enhance the Raman signal intensity. Based on its excellent SERS enhancement performance, this gold-silver alloy nanoflower SERS substrate is applied to the immune recognition-mediated sandwich detection mode to achieve highly sensitive and accurate detection of occult blood markers in fecal samples. The preparation process of this invention is simple, low-cost, and requires no complex equipment. The resulting SERS substrate has strong stability and excellent repeatability, which can meet the needs of early screening and rapid on-site detection of digestive tract diseases.

This invention discloses a macro- and micro-image feature fusion and classification method based on the identification of the same lesion. It not only adaptively designs spatial and channel attention mechanisms to extract macro- and micro-features respectively, taking into account the visual characteristics of different modalities, but also introduces a cross-modal feature fusion and bias reconciliation mechanism. By quantifying cross-modal semantic consistency, it dynamically extracts and suppresses systematic bias components. At the same time, it performs deep difference reconciliation and residual compensation on bimodal features in local and global receptive fields, breaking through the performance bottleneck of traditional multimodal direct stitching. It achieves efficient and accurate fusion of macro-morphological and micro-cellular features, and can perform intelligent identification of digestive tract diseases based on the macro- and micro-features and their fused features. It has stronger robustness and accuracy, significantly improves the accuracy of lesion identification, and helps doctors obtain more accurate diagnostic results.