A realistic model of human finger was constructed based on high-resolution MRI image. The model was used to simulate transmission and reflectance images of healthy and inflamed joints by different illumination geometries. Two optical windows were determined, optimal illumination identified and characteristic inflammation features found. The simulation results were compared to experimental images and good agreement was obtained.
COBISS.SI-ID: 3386980
We constructed a laser profilomer for measuring 3D surface of imaged objects. We developed an algorithm for correction of curvature and height artefacts in spectral images using the 3D information. We demonstrated that the approach almost completely eliminated the artifacts on surfaces with inclination up to 60 deg.
COBISS.SI-ID: 3383908
Mouse models of peritoneal inflammation were imaged by hyperspectral system. Fast algorithm for extraction of physiologic and morphologic parameters from spectral images was developed. Based on the extracted properties we were able to classify inflamed and control animals. The results agreed with the histology of the imaged tissues.
COBISS.SI-ID: 3414372
A robust and stable tissue phantoms were developed from silicone substrate. Phantoms with different absorption and scattering properties were fully optically characterized. We demonstrated that such phantoms are suitable for an optical system validation.
COBISS.SI-ID: 19976451
In a model of human finger we simulated reflection and transmission of light in the finger tissues. Based on the simulation results we determined spectral bands where characteristic features of joint inflammation are present. These results guided us at development of multispectral imaging system for human joints.
COBISS.SI-ID: 3180900