Automatic detection of basal cell carcinoma by hyperspectral imaging

Abstract

The purpose of this study was to test the ability of hyperspectral imaging (HSI) combined with unsupervised anomaly detectors to automatically differentiate basal cell carcinoma (BCC) from normal skin. Hyperspectral images of the face of a female patient with a BCC of the lower lip were acquired using a visible/near-infrared HSI system and two anomaly detection algorithms (Reed-Xiaoli and Reed-Xiaoli/Uniform Target hybrid anomaly detectors) were used to detect pathological tissue from normal skin. The results revealed that the receiver operating characteristic curve of the Reed-Xiaoli/Uniform Target hybrid detector was higher than that of the Reed-Xiaoli detector in the range of false positive rates between 0 and 0.8. The area under curve values were good (0.7074 and 0.8607, respectively) with Reed-Xiaoli/Uniform Target hybrid detector performing better. In conclusion, HSI combined with either of two anomaly detectors can play a promising role in the automated screening of BCC.

Single probe light reflectance spectroscopy and parameter spectrum feature extraction in experimental skin cancer detection and classification

“…Specifically, evaluation of the performance of the proposed hybrid strategy with a larger number of subjects may ultimately lead to the development of an algorithm for automated cancer identification and screening with the use of deep learning and machine learning methods. Regarding our future plans, further development of the method is warranted including application of an AI approach for the development of automated tissue classification methods which can later be applied to the assessment of tumor margins based on the hyperspectral imaging concept presented here [20,32,[44][45][46][47][48][49]. One way to improve our screening performance is to separately obtain the scattering coefficient from the absorption coefficient by utilizing the approach described in ref.…”

Hyperspectral imaging for early diagnosis of diseases: A review

“…Following a novel approach, Q. Wang et al, 2020 proposed a deep convolutional network (3D CNN) for the segmentation of melanoma hyperspectral pathology images. In yet new research, automatic differentiation of basal cell carcinoma (BCC) from normal skin using HSI was conducted by Calin & Parasca, 2022, using two unsupervised anomaly detectors, namely ‐ the RXD (Reed‐Xiaoli anomaly detector) and the hybrid RXD‐UTD (Reed‐Xiaoli Uniform Target detector).…”

“…Noise removal from hyperspectral images can be done using moving average filter(M. H. Aref et al, 2020), inverse minimum noise fraction (Calin & Parasca, 2022) and 2nd and 1st derivative computation (Hornberger et al, 2020), (Hu et al, 2019). Generally, de‐noising is done simultaneously with normalization using white and dark references.…”

“…Calin & Parasca, 2022 in their study proposed an automatic differentiation of BCC (basal cell carcinoma) of normal skin using HSI. The images were taken using a VNIR HSI system from a female patient face who had been diagnosed with BCC in the lower lip.…”

“…Moreover, hyperspectral imaging is also helpful in the non-invasive identification and diagnosis of diseases like brain tumour detection, skin cancer, colon cancer, and many more. HSI measures the spectral signatures of different wavelengths reflected and absorbed by the light, allowing simultaneous measurement of different tissue constituencies and their spatial distribution from each hyperspectral image pixel (Calin et al, 2014).…”