Photoacoustic Remote Sensing (PARS®) is a non-contact, label-free imaging modality that provides optical absorption contrast in biological tissues. Images are formed by raster-scanning over a target. A time-domain signal is collected at each point, representing initial pressure-induced via the photoacoustic effect. Conventionally, only the amplitude of the time-domain signals is considered to estimate pixel values, disregarding the rich temporal information present in the signals. For instance, the signal shape carries information, which may be related to specific biological structures. In this work, clustering based on signal shape is explored, followed by feature extraction, enabling the virtual labeling of PARS images.
|