Репозиторий Университета

© 2018 SPIE. We have proposed a wavelet-domain de-noising technique for imaging of human brain malignant glioma by optical coherence tomography (OCT). It implies OCT image decomposition using the direct fast wavelet transform, thresholding of the obtained wavelet spectrum and further inverse fast wavelet transform for image reconstruction. By selecting both wavelet basis and thresholding procedure, we have found an optimal wavelet filter, which application improves differentiation of the considered brain tissue classes-i.e. malignant glioma and normal/intact tissue. Namely, it allows reducing the scattering noise in the OCT images and retaining signal decrement for each tissue class. Therefore, the observed results reveals the wavelet-domain de-noising as a prospective tool for improved characterization of biological tissue using the OCT.

© COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only. We have performed the in vitro terahertz (THz) spectroscopy of human brain tumors. In order to fix tissues for the THz measurements, we have applied the gelatin embedding. It allows for preserving tissues from hydration/dehydration and sustaining their THz response similar to that of the freshly-excised tissues for a long time after resection. We have assembled an experimental setup for the reflection-mode measurements of human brain tissues based on the THz pulsed spectrometer. We have used this setup to study in vitro the refractive index and the amplitude absorption coefficient of 2 samples of malignant glioma (grade IV), 1 sample of meningioma (grade I), and samples of intact tissues. We have observed significant differences between the THz responses of normal and pathological tissues of the brain. The results of this paper highlight the potential of the THz technology in the intraoperative neurodiagnosis of tumors relying on the endogenous labels of tumorous tissues.