[Noninvasive medical imaging system for tissue classification using RGB LED and micro-spectroscopy].

As skin is the exterior organ of human body, cosmetic industry advances year by year. To reveal the details of skin tissue, threedimensional medical imaging is required. Based on the idea of "readout instead of write", a new scheme named spectral classification imaging (SCI) is proposed in the present study to reduce the invasiveness by applying the reflection spectra of the sample points for three-dimensional medical imaging. Broad-band light source and the spectrometer were employed to collect the spectra curves of scanned region, which were classified into several tissue types by their cross-correlations. A colorful tissue tomography can finally be obtained by filling in each image pixel the color indicating the corresponding tissue type. The lateral/longitudinal resolutions and penetration depth were analyzed to characterize the SCI system. The lateral resolution is based on the source's diffraction limit, the longitudinal resolution is by its depth-of-focus, and the penetration depth is equivalent to its skin depth. The imaging results of an amethyst of 0.6 mm (chi-direction) x 0.6 mm (y-direction) with a total of 120 x 120 pixels per frame and a guppy fish of 3.2 mm (chi-direction) x 2.4 mm (y-direction) of 160 x 120 pixels, are presented to show the image quality. The effects of the cross-correlation coefficient and the number of source wavelengths on the imaging results were explored. The value of cross-correlation threshold determines the required time for imaging, the resulted number of tissue groups, and the variety of tissue colors in the imaging result. Owing to its virtual noninvasiveness and easy configuration, the SCI system is highly promising for practical uses. RGB LEDs possess merits of broad bandwidth, low cost, long lifetime, small volume, and are ready to be integrated into a multi-color source module. Replacing the wide-band light source and the spectrometer module with a composite RGB LED with discrete wavelengths and a micro-spectrometer for spectra retrieval, the system has great potential to be minimized as a hand-held product for noninvasive medical imaging. It leads to reduced use of non-eco-friendly cosmetics and extended advance of cosmetic dermatology.