Comprehensive Image-processing Technique for Hemoglobin Analysis and Applications in Laparoscopic System Integration

Surgical cameras and endoscopes provide important imaging information about a patient’s anatomy and pathology, but most of the data are limited to 2D textured representations of visible tissue surfaces. Conversely, research is progressing in the information acquisition field by using spectrum data to obtain more detailed disease diagnostic support information. Spectroscopic information reflects the chemical and physical information of substances. Research is expanding from conventional single-point spectroscopic information to acquisition of two-dimensional spectroscopic information, called hyperspectral data. However, this type of information typically takes a relatively long time to acquire and analyze, which has limited its practical use. Particularly, compensation for the continuous motions of the camera and tissue occurring during acquisition of numerous images as well as the analysis time slow delivery of real-time information. Due to these disadvantages, the use of analyzed hyperspectral data information in medicine has been limited to the research stage. This study aimed to use this system to visualize blood phantoms and biological information from small animals. We developed a 3-band spectral imaging method for measuring relative blood concentration and oxygen saturation, and modified white light images. Real-time information display is possible by limiting the target information and speeding up image acquisition and processing. We also developed an existing red‐green‐blue (RGB) wavelength-band digital imaging system, in which an optical filter is used to limit the wavelength range to effectively capture the spectral changes associated with oxygen saturation and concentration. Since the three-wavelength information is in the wavelength range of the RGB channels of the image sensor, single-frame data capture was achieved. An analytical method for 3-band spectral data was also developed, which contributed to fast image processing. Further development of this technology can eliminate the disadvantages associated with conventional spectroscopic imaging and analysis. Consequently, we expect that it will contribute to practical use of spectral information in the medical field. The method developed in this study can be applied to general purpose cameras and is highly practical as a display technology for diagnostic support information.