Subcutaneous Vascular Structure Enhancement: A Tri-Modal 850 nm NIR Image-Processing Pipeline for Palm-Vein Visualization

Near-infrared (NIR) images of palms are affected by non-uniform illumination, sensor noise, skin texture, and capture variation. This article follows one clear, complete grayscale capture from the public SASH-VPV sampler through the proposed tri-modal processing pipeline. Each operation is shown as a large side-by-side comparison of its direct input and output, using the same source image throughout. The sequence includes foreground masking, bilateral denoising, CLAHE, background estimation, dark-ridge response, Gabor, Frangi, MPC, fusion, adaptive thresholding, confidence gating, and morphology. Original design values are retained where specified; additional implementation settings are identified as illustrative. The outputs also enhance palm creases and hand boundaries, so these figures demonstrate processing transitions, not validated vessel segmentation. The result is a candidate feature mask for research and prototyping, not a clinical or biometric result.