Abstract-Methods for registering and calibrating in vivo terahertz images of cutaneous burn wounds

Priyamvada Tewari, James Garritano, Neha Bajwa, Shijun Sung, Haochong Huang, Dayong Wang, Warren Grundfest, Daniel B. Ennis, Dan Ruan, Elliott Brown, Erik Dutson, Michael C. Fishbein, and Zachary Taylor

Fig. 1 Burn zones organize as concentric like shells in the depth are hypothesized to appear as projections of the shells on the surface. Skin layers denoted A, B, C and damage zones denoted 1, 2, 3, 4. (a) Cross-sectional view: damage from superficial thickness wounds are limited primarily to the stratum corneum and epidermis, partial thickness burn wounds extend into the epidermis, and full thickness wounds extend into deep dermis. (b) Isometric view providing orientation between (a) and (c). (c) Top down view: The damage zones, and thus variations in surface TWC should present as regions concentric with the center.

https://www.osapublishing.org/boe/abstract.cfm?uri=boe-10-1-322

A method to register THz and visible images of cutaneous burn wounds and to calibrate THz image data is presented. Images of partial and full thickness burn wounds in 9 rats were collected over 435 mins. = 7.25 hours following burn induction. A two-step process was developed to reference the unknown structure of THz imaging contrast to the known structure and the features present in visible images of the injury. This process enabled the demarcation of a wound center for each THz image, independent of THz contrast. Threshold based segmentation enabled the automated identification of air (0% reflectivity), brass (100% reflectivity), and abdomen regions within the registered THz images. Pixel populations, defined by the segmentations, informed unsupervised image calibration and contrast warping for display. The registered images revealed that the largest variation in THz tissue reflectivity occurred superior to the contact region at ~0.13%/min. Conversely the contact region showed demonstrated an ~6.5-fold decrease at ~0.02%/min. Exploration of occlusion effects suggests that window contact may affect the measured edematous response.

© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement