Annu Int Conf IEEE Eng Med Biol Soc. 2021 Nov;2021:3605-3608. doi: 10.1109/EMBC46164.2021.9629512.
Pigmented skin lesions (PSL) are prevalent in Asian populations and their gross pathology remains a manual, tedious task. Hyper-spectral imaging (HSI) is a non-invasive non-ionizing acquisition technique, allowing malignant tissue to be identified by its spectral signature. We set up a hyper-spectral imaging (HSI) system targeting cancer margin detection of PSL. Because classification among PSL is achieved via comparison of spectral signatures, appropriate calibration is necessary to ensure sufficient data quality. We propose a strategy for system building, calibration and pre-processing, under the requirements of fast acquisition and wide field of view. Preliminary results show that the HSI-based system is able to effectively resolve reflectance signatures of ex-vivo tissue.Clinical Relevance-The imaging system proposed in this study can recover reflectance spectra from PSL during gross pathology, providing a wide imaging area.
色素性皮肤病变（PSL）在亚洲人群中很常见，其大体病理学仍然是一项手动、繁琐的任务。高光谱成像（HSI）是一种非侵入性的非电离采集技术，允许通过其光谱特征来识别恶性组织。我们建立了一个针对 PSL 癌症边缘检测的高光谱成像（HSI）系统。因为 PSL 之间的分类是通过光谱特征的比较来实现的，所以需要适当的校准来确保足够的数据质量。根据快速采集和宽视场的要求，我们提出了一种系统构建、校准和预处理的策略。初步结果表明，基于 HSI 的系统能够有效地解析离体组织的反射率特征。临床相关性-本研究中提出的成像系统可以在大体病理学过程中从 PSL 中恢复反射光谱，提供更宽的成像区域。
