Abstract
We present an instrument based on commodity embedded hardware, that implements an automatic procedure for early skin-cancer screening using dynamic thermal imaging. The procedure leverages image segmentation in the visible range and real-time multimodal registration to compute the temperature recovery curve (TRC) of suspicious skin lesions using thermal infrared video. The instrument implements two algorithms that infer the malignancy of the lesion from the computed TRCs. The first algorithm assumes that the TRCs are deterministic and infers the malignancy from the distance between the TRC of the suspicious lesion and its surrounding skin, which is assumed to be healthy tissue. The second algorithm models the TRC of the lesion as a random process and uses detection theory to statistically infer its malignancy from the eigenfunctions and corresponding eigenvalues of its covariance function. We built a prototype of the instrument using a Raspberry Pi 3 model B+ board, which acquires a visible-range image of the lesion at the beginning of the procedure and performs image segmentation in 62ms. Operating on a 400×400-pixel region-of-interest within the infrared video, the board performs frame-by-frame multimodal image registration and generates the TRCs in real time at more than 37 frames per second, thus eliminating the need to store video data for off-line processing. The statistical detection algorithm, which yields the best results, runs in 1.07s at the end of the procedure, and achieves a sensitivity of 98% and a specificity of 95% on a dataset of 116 volunteer subjects.
Highlights
The incidence of skin cancer has increased over the past decades and is the most commonly-diagnosed cancer in the United States, where the number of new melanoma cases has increased annually by 53% [1]
It is estimated that melanoma occurs in only 4% of all skin cancer cases, it is the most lethal type, being responsible for about 75% of the deaths caused by skin cancer in the United States [1]
Current detection procedures include the evaluation of a dermatologist through the ABCDE method [5], in which the physician searches for irregularities in the suspected lesion in terms of its asymmetry, border, color, diameter, and evolution [6]
Summary
The incidence of skin cancer has increased over the past decades and is the most commonly-diagnosed cancer in the United States, where the number of new melanoma cases has increased annually by 53% [1]. More recent studies have reported results of the use of deep-learning algorithms to predict the malignancy of skin cancer lesions; for example, Dascalu and David [28] achieved a sensitivity of 0.917 and a specificity of 0.418 by using using the same Google CNN architecture to extract features from the images, which are modeled and processed as a sound signal. Brinker et al [29] use a CNN architecture, which achieves a sensitivity of 0.865 and a specificity or 0.875, outperforming 136 out of 157 recruited dermatologists Another approach that has been investigated to aid in the detection of skin cancer in its early state is the use of infrared (IR) thermal imaging [30].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
