ABSTRACT
The physiological characteristics such as dark skin tone of the patients lead to the problem of peripheral difficult venous access (PDVA). It occurs during the intravenous (IV) catheterization process for drugs delivery or blood sampling and results in pain and trauma for the patients. In order to overcome the PDVA problem, this chapter discusses the optimization of illumination band for near infrared (NIR) imaging modality using hyperspectral image analysis. A hyperspectral dataset of 252 subjects from different skin tone classes is formed and analysed using principal component analysis (PCA), linear discriminant analysis (LDA) and objective hyperspectral image quality assessment. On the basis of skin tone, the hyperspectral images are classified into four classes: fair, light brown, dark brown and dark. The conclusion drawn from this analysis is that the illumination wavelength band of 800–850 nm is optimum for the better venous image quality for all skin tone subjects. This illumination band can be used in venous imaging systems to get better venous images despite the difference in skin tones of the patients.
