Stereoscopic image data in the near-infrared (NIR) spectral range is used for the precise detection of subcutaneous veins. By using neural networks, venous structures can be reliably segmented even under difficult conditions.
A three-dimensional reconstruction of the venous network is created on the basis of the segmented image data. The combination of multimodal imaging, AI-supported segmentation and 3D modeling enables an exact, spatial representation of the vascular structure.
Such technologies form the basis for intelligent assistance systems in medical applications - for example to support punctures or to plan robotic access.
Visualization and processing of forearm veins using a stereoscopic camera system. a) Original images of a forearm, taken with a specially developed stereo camera system. b) Visualization of the veins using NIR filters. c) & d) Image processing for edge detection and segmentation of the veins using masking. e) 3D reconstruction of the vein tree from two different perspectives.