Stitching technique applied to UV intrinsic fluorescence imaging in an in vivo wound-healing model

Chronic ulcers are skin wounds, which have severe repercussions for patients regarding mobility restriction and economic problems. Nowadays, no technique allows to evaluate the evolution of ulcers' healing process in clinical practice consistently. UV fluorescence excitation photography can provide spatial and temporal information on molecular and structural changes, providing an objective means for evaluating the healing process. In this work, the UV intrinsic fluorescence from the reepithelization process was imaged in an in vivo wound animal model. Scan imaging process and imaging misalignments were emulated. We evaluated SURF and RANSAC algorithms for stitching low-intensity and low contrast images. The evaluated algorithms could identify up to 16 common characteristics in contiguous images with a 20% overlap area. The maximum error found in the stitching process in an 18-day healing period was 1.69%.