Hyperspectral Reflectance Imaging Research Articles

Mapping of fatty acids composition in shelled almonds analysed in bulk using a Hyperspectral Imaging system

The determination of the fatty acid profile in food products is an important issue, as it serves as a guide to consumers who wish to follow healthy diets. Hyperspectral Imaging (HSI), permits rapid, non-destructive quality evaluation in foods by integrating the spatial dimension of the composition distribution. The aim of this research was to measure the fatty acid profile of almonds using HSI in 149 samples of shelled sweet and bitter almonds. In addition, we analysed the inter- and intra-kernel distribution of fatty acids for both type of almonds. Shelled sweet and bitter almonds were scanned in bulk by reflectance HSI (946.6–1648.0 nm) and then analysed by gas chromatography to determine their fatty acid composition. Next, we built quantitative prediction models using Partial Least Squares (PLS) regression and tested two validation strategies — mean spectrum and pixel-by-pixel. The developed HSI calibration models showed a good performance when quantifying oleic and linoleic acids, while the models developed could be used for screening purposes for the rest of the fatty acids analysed and for the oleic to linoleic ratio. The results obtained confirm that HSI can be considered a promising approach for estimating fatty acids and their inter- and intra-kernel distribution.

Identification of Corn Seeds with Different Freezing Damage Degree Based on Hyperspectral Reflectance Imaging and Deep Learning Method

Seed freezing damage is an agricultural disaster, which has a great impact on seed quality assurance. The feasibility of combining hyperspectral imaging with deep convolutional neural network (DCNN) to classify different freeze-damaged corn seeds was studied in this paper. At first, the hyperspectral images of corn seeds subjected to five different freezing temperatures at 400–1000 nm were acquired, and then the average spectra were extracted from the region of embryo hyperspectral images over the wavelength range of 450–979 nm. Next, four models (K nearest neighbors (KNN), support vector machine (SVM), extreme learning machine (ELM), and DCNN) were developed for five-category (5 frozen conditions) and four-category (“no freezing,” “slight freezing,” “moderate freezing,” and “severe freezing”) classifications, and the values of the evaluation indexes (accuracy, sensitivity, specificity, and precision) were calculated for comparison. The results show that DCNN model had the most satisfactory result with accuracy rates of 100% (training set), 96.9% (validation set), and 97.5% (testing set) for five-category classification, with accuracy rates of 100% (training, validation, and testing set) for four-category classification, and DCNN model also had the best performance in the evaluation indexes. At last, the visual classification map was generated according to the results of DCNN. It shows that hyperspectral imaging and DCNN can provide a novel method to detect the freezing damage of corn seeds quickly and inexpensively.

Study on the Relations between Hyperspectral Images of Bananas (Musa spp.) from Different Countries, Their Compositional Traits and Growing Conditions.

Bananas are some of the most popular fruits around the world. However, there is limited research that explores hyperspectral imaging of bananas and its relationship with the chemical composition and growing conditions. In the study, the relations that exist between the visible near-infrared hyperspectral reflectance imaging data in the 400–1000 nm range of the bananas collected from different countries, the compositional traits and local growing conditions (altitude, temperature and rainfall) and production management (organic/conventional) were explored. The main compositional traits included moisture, starch, dietary fibre, protein, carotene content and the CIE L*a*b* colour values were also determined. The principal component analysis showed the preliminary separation of bananas from different geographical origins and production systems. The compositional and spectral data revealed positively and negatively moderate correlations (r around ±0.50, p < 0.05) between the carotene, starch content, and colour values (a*, b*) on the one hand and the wavelength ranges 405–525 nm, 615–645 nm, 885–985 nm on the other hand. Since the variation in composition and colour values were related to rainfall and temperature, the spectral information is likely also influenced by the growing conditions. The results could be useful to the industry for the improvement of banana quality and traceability.

Corn seed variety classification based on hyperspectral reflectance imaging and deep convolutional neural network

Variety purity is an important indicator in seed quality detection. Different varieties of corn seeds may be mixed in the growth and development process, which affects the growth and yield of the seeds. Thus, it is necessary to find a fast and non-destructively method to detect the purity. In this paper, the feasibility of combining hyperspectral imaging with deep convolutional neural network (DCNN) was studied to classify four corn seed varieties. Firstly, the average spectra from the region of seed in endosperm side hyperspectral images over the wavelength range of 450–979 nm were extracted. Secondly, the performances of three models were compared, including DCNN, K nearest neighbors (KNN) and support vector machine (SVM). DCNN model has the 100% training accuracy rate, 94.4% testing accuracy rate and 93.3% validation accuracy rate, and outperforms KNN and SVM models in most cases. DCNN model also had the best performance in evaluation indexes (sensitivity, specificity and precision). Finally, the visual classification map was generated according to the results of DCNN. Results show that DCNN can be adopted in spectral data analysis for the variety classification of corn seed; and the classification performance can be improved effectively.

Combination of noninvasive imaging techniques to characterize pigments in Buddhist thangka paintings

Scanning macro‐XRF, XRF point measurements and hyperspectral reflectance imaging in the visible and near‐infrared range were combined to characterize the pigments of a Tibetan painting. The complementary use of the different investigation techniques allowed us to (a) clearly identify the pigments used by the artist, (b) strengthen the few data present in historical and contemporary sources and (c) provide data necessary for any accurate conservation treatment of the painting. The proposed methodology allows an in‐depth characterization of the materials used in thangkas, with the identification of all the pigments used by Tibetan artists, as shown by the analyses of reference samples. In the present study traditional pigments, commonly cited in historical sources, were identified such as vermilion, minium, orpiment and azurite, but also more unusual compounds such as brochantite and antlerite (in blue/green areas). Elemental distribution images confirmed that some parts of the thangka were restored, without taking into account the original pigments or the original design. Underlying annotations and a carbon black under‐drawing were highlighted by the application of near‐infrared hyperspectral imaging technique, which proved to be particularly useful for the study of such paintings.

Hyperspectral reflectance imaging to classify lettuce varieties by optimum selected wavelengths and linear discriminant analysis

Lettuce (Lactuca sativa L.) has a wide variation of pigment classes and content in its leaves, with a variety of colors, textures, and sizes, which impose difficulties for their classification. Currently, the search for varieties with higher levels of antioxidant compounds has led the consumer to seek those with a purple color. However, there are no classification systems on the part of the distributor sector that allow the product delivered to the market to be directed. Thus, the use of remote sensing has become essential for better classification, providing nondestructive, fast, and precise measurement. Based on this, the aim of this study was to examine the potential of hyperspectral imaging for classifying different lettuce varieties through multivariate statistical techniques. Hyperspectral images in 825 visible infrared wavelengths (from 350 to 1000 nm) were acquired for eleven varieties with different foliar pigments (chlorophyll and anthocyanin). Analysis of pigments, analysis of tissues by optical microscopy, principal component analysis (PCA) and linear discriminant analysis (LDA), based on STEPWISE wavelengths selection were made. The results showed that there are differences in the leaf thickness and the anthocyanin overlap under chlorophyll in purple varieties, resulting in different spectral responses. The PCA explained 83% of variability in the data in the first two components, separate clusters were formed between the purple and green varieties. The STEPWISE procedure selected 31 wavelengths: 12 in the red, 9 in the green, 1 in the blue, and the remainder in the near infrared spectrum. The LDA models achieved a classification success rate of 81.04%. The results show that the use of hyperspectral imaging has great potential to distinguish among lettuce varieties. In the future, these results can support the development of a quickly analysis with semi-automated procedures in a repeatable, fast, and objective manner that can be integrated in farms and distribution centers.

REFLECTANCE RECONSTRUCTION OF HYPERSPECTRAL IMAGE BASED ON GAUSSIAN SURFACE FITTING

Abstract. Different from the field of remote sensing, artificial lights are often utilized as the energy source for spectral imaging in the ground hyperspectral applications. The kind of double-spot light source is widely adopted in some large scale ground hyperspectral applications. However, it is hard to reach a satisfied lighting without difference in light intensity in many cases although the lamps are tuned carefully. Therefore, a reflectance calibration of hyperspectral imaging based on the data of diffuse reflectance standard and Gaussian surface fitting is proposed in this paper. The purpose is to improve the reconstruction accuracy of hyperspectral reflectance image by minimized the error caused by the uneven illumination of artificial light source. The method has a higher accuracy than traditional one according to the experiment results.

An alternative approach to mapping pigments in paintings with hyperspectral reflectance image cubes using artificial intelligence

Spectral imaging modalities, including reflectance and X-ray fluorescence, play an important role in conservation science. In reflectance hyperspectral imaging, the data are classified into areas having similar spectra and turned into labeled pigment maps using spectral features and fusing with other information. Direct classification and labeling remain challenging because many paints are intimate pigment mixtures that require a non-linear unmixing model for a robust solution. Neural networks have been successful in modeling non-linear mixtures in remote sensing with large training datasets. For paintings, however, existing spectral databases are small and do not encompass the diversity encountered. Given that painting practices are relatively consistent within schools of artistic practices, we tested the suitability of using reflectance spectra from a subgroup of well-characterized paintings to build a large database to train a one-dimensional (spectral) convolutional neural network. The labeled pigment maps produced were found to be robust within similar styles of paintings.

Pablo Picasso’s Mother and Child by the Sea (1902): report on the hyperspectral near-infrared reflectance imaging survey of Picasso’s newspaper use

Picasso reused the canvas of Mother and Child by the Sea (1902) several times. A hyperspectral near-Infrared (960–1680 nm) reflectance imaging survey of the painting revealed newsprint text from Le Journal of January 1902. Newsprint from Le Journal was identified as well in two other paintings, Jaume Sabartes with Pince-Nez (1901) and Portrait of Mateu Fernandez de Soto (1901). Picasso transported these two portraits from Paris to Barcelona along with Mother and Child by the Sea. In all cases, the newsprint letters appear reversed. This presumably was the result of newspaper being applied, for the purpose of transport, to the surface of paintings that were not entirely dry. Inspection of the painting and the infrared hyperspectral images suggest that once this ‘ink transfer’ occurred it appears that Picasso left the newsprint text in place when he went on to complete the painting. Although Picasso’s use of newspaper in his compositions is not unusual, further investigation of the presence of newsprint letters in Blue Period works is of interest. Moreover, the survey’s identification of the particular date of the newspaper in question provides important information regarding the timing of Picasso’s movement from Paris to Barcelona, and concerning his repainting and completion of each work.

Classification of Hyperspectral Reflectance Images With Physical and Statistical Criteria

A classification method of hyperspectral reflectance images named CHRIPS (Classification of Hyperspectral Reflectance Images with Physical and Statistical criteria) is presented. This method aims at classifying each pixel from a given set of thirteen classes: unidentified dark surface, water, plastic matter, carbonate, clay, vegetation (dark green, dense green, sparse green, stressed), house roof/tile, asphalt, vehicle/paint/metal surface and non-carbonated gravel. Each class is characterized by physical criteria (detection of specific absorptions or shape features) or statistical criteria (use of dedicated spectral indices) over spectral reflectance. CHRIPS input is a hyperspectral reflectance image covering the spectral range [400–2500 nm]. The presented method has four advantages, namely: (i) is robust in transfer, class identification is based on criteria that are not very sensitive to sensor type; (ii) does not require training, criteria are pre-defined; (iii) includes a reject class, this class reduces misclassifications; (iv) high precision and recall, F 1 score is generally above 0.9 in our test. As the number of classes is limited, CHRIPS could be used in combination with other classification algorithms able to process the reject class in order to decrease the number of unclassified pixels.

Non-Destructive Detection of Bone Fragments Embedded in Meat Using Hyperspectral Reflectance Imaging Technique.

Meat consumption has shifted from a quantitative to a qualitative growth stage due to improved living standards and economic development. Recently, consumers have paid attention to quality and safety in their decision to purchase meat. However, foreign substances which are not normal food ingredients are unintentionally incorporated into meat. These should be eliminated as a hazard to quality or safety. It is important to find a fast, non-destructive, and accurate detection technique of foreign substance in the meat processing industry. Hyperspectral imaging technology has been regarded as a novel technology capable of providing large-scale imaging and continuous observation information on agricultural products and food. In this study, we considered the feasibility of the short-wave near infrared (SWIR) hyperspectral reflectance imaging technique to detect bone fragments embedded in chicken meat. De-boned chicken breast samples with thicknesses of 3, 6, and 9-mm and 5 bone fragments with lengths of about 20–30-mm are used for this experiment. The reflectance spectra (in the wavelength range from 987 to 1701-nm) of the 5 bone fragments embedded under the chicken breast fillet are collected. Our results suggested that these hyperspectral imaging technique is able to detect bone fragments in chicken breast, particularly with the use of a subtraction image (corresponding to image at 1153.8-nm and 1480.2-nm). Thus, the SWIR hyperspectral reflectance imaging technique can be potentially used to detect foreign substance embedded in meat.

Non-Destructive Detection of Strawberry Quality Using Multi-Features of Hyperspectral Imaging and Multivariate Methods.

Soluble solid content (SSC), pH, and vitamin C (VC) are considered as key parameters for strawberry quality. Spectral, color, and textural features from hyperspectral reflectance imaging of 400–1000 nm was to develop the non-destructive detection approaches for SSC, pH, and VC of strawberries by integrating various multivariate methods as partial least-squares regression (PLSR), support vector regression, and locally weighted regression (LWR). SSC, pH, and VC of 120 strawberries were statistically analyzed to facilitate the partitioning of data sets, which helped optimize the model. PLSR, with spectral and color features, obtained the optimal prediction of SSC with determination coefficient of prediction (Rp2) of 0.9370 and the root mean square error of prediction (RMSEP) of 0.1145. Through spectral features, the best prediction for pH was obtained by LWR with Rp2 = 0.8493 and RMSEP = 0.0501. Combination of spectral and textural features with PLSR provided the best results of VC with Rp2 = 0.8769 and RMSEP = 0.0279. Competitive adaptive reweighted sampling and uninformative variable elimination (UVE) were used to select important variables from the above features. Based on the important variables, the accuracy of SSC, pH, and VC prediction both gain the promotion. Finally, the distribution maps of SSC, pH, and VC over time were generated, and the change trend of three quality parameters was observed. Thus, the proposed method can nondestructively and accurately determine SSC, pH, and VC of strawberries and is expected to design and construct the simple sensors for the above quality parameters of strawberries.

Detection and community-level identification of microbial mats in the McMurdo Dry Valleys using drone-based hyperspectral reflectance imaging

AbstractThe reflectance spectroscopic characteristics of cyanobacteria-dominated microbial mats in the McMurdo Dry Valleys (MDVs) were measured using a hyperspectral point spectrometer aboard an unmanned aerial system (remotely piloted aircraft system, unmanned aerial vehicle or drone) to determine whether mat presence, type and activity could be mapped at a spatial scale sufficient to characterize inter-annual change. Mats near Howard Glacier and Canada Glacier (ASPA 131) were mapped and mat samples were collected for DNA-based microbiome analysis. Although a broadband spectral parameter (a partial normalized difference vegetation index) identified mats, it missed mats in comparatively deep (&gt; 10 cm) water or on bouldery surfaces where mats occupied fringing moats. A hyperspectral parameter (B6) did not have these shortcomings and recorded a larger dynamic range at both sites. When linked with colour orthomosaic data, B6 band strength is shown to be capable of characterizing the presence, type and activity of cyanobacteria-dominated mats in and around MDV streams. 16S rRNA gene polymerase chain reaction amplicon sequencing analysis of the mat samples revealed that dominant cyanobacterial taxa differed between spectrally distinguishable mats, indicating that spectral differences reflect underlying biological distinctiveness. Combined rapid-repeat hyperspectral measurements can be applied in order to monitor the distribution and activity of sentinel microbial ecosystems across the terrestrial Antarctic.

Use of Spectroscopic Techniques for a Rapid and Non-Destructive Monitoring of Thermal Treatments and Storage Time of Sous-Vide Cooked Cod Fillets.

In this work, the potential of spectroscopic techniques was studied to investigate heat-induced changes occurring during the application of thermal treatments on cod (Gadus morhua L.) fillets. Vacuum-packed samples were thermally treated in a water bath at 50, 60, 70 and 80 °C for 5 and 10 min, and further stored for one, four, and eight days at 4 ± 1 °C before analysis. Several traditional (including cooking loss, drip loss, texture, protein solubility, protein oxidation, and color) and spectroscopic (fluorescence and diffuse reflectance hyperspectral imaging) measurements were conducted on the same samples. The results showed a decrease in fluorescence intensity with increasing cooking temperature and storage time, while the impact of cooking time was only noticeable at low temperatures. Diffuse reflectance data exhibited a decrease in absorbance, possibly as a result of protein denaturation and increased scattering at higher cooking temperatures. Both fluorescence and diffuse reflectance data were highly correlated with color parameters, whereas moderate correlations were observed with most other traditional parameters. Support vector machine models performed better than partial least square ones for both classification of cod samples cooked at different temperatures and in prediction of the cooking temperature. The best classification result was obtained on fluorescence data, achieving an accuracy of 92.5%, while the prediction models resulted in a root mean square error of prediction of cooking temperature lower than 5 °C. Overall, the classification and prediction models showed good results, indicating that spectroscopic techniques, especially fluorescence hyperspectral imaging, have a high potential for monitoring thermal treatments in cod fillets.

Performance of Fluorescence and Diffuse Reflectance Hyperspectral Imaging for Characterization of Lutefisk: A Traditional Norwegian Fish Dish.

Lutefisk is a traditional Norwegian fish dish made from dried fish, such as cod or other whitefish. In Norway and other Nordic countries, lutefisk is considered among the most popular dishes served during Christmas or other festive occasions. However, to date, little attention has been paid to this product, and available research on the quality, processing, and chemistry of lutefisk is still limited. The quality of this very delicate product, with a high pH value, depends on many factors, such as the initial quality of raw materials (stockfish), the quantity of lye used during the preparation process of lutefisk, and time during soaking in the lye and water, among others, making it challenging to both optimize processing and monitor the quality of lutefisk. In this study, four commercially available lutefisk brands (labelled as A, B, C, and D) were characterized using two online spectroscopic techniques, namely fluorescence and diffuse reflectance hyperspectral imaging, implemented on conveyor belts to mimic industrial applications. The samples were also analyzed by the use of an offline laboratory instrument based on visible/near infrared diffuse reflectance spectroscopy. Three traditional measurements, including texture, water content, and pH, were also conducted on the same samples. Supervised classification PLS-DA models were built with each dataset and relationships between the spectroscopic measurements and the traditional data were investigated using canonical correlations. The spectroscopic methods, especially fluorescence spectroscopy, demonstrated high performance for the discrimination between samples of the different brands, with high correlations between the spectral and traditional measurements. Although more validations of the results of this study are still required, these preliminary findings suggest that the destructive, laborious, and time-consuming traditional techniques can be replaced by rapid and nondestructive online measurements based on hyperspectral imaging used in fluorescence or diffuse reflectance mode.

Disentangling the effects of doping, strain and disorder in monolayer WS2 by optical spectroscopy

Monolayers of transition metal dichalcogenides (TMdC) are promising candidates for realization of a new generation of optoelectronic devices. The optical properties of these two-dimensional materials, however, vary from flake to flake, or even across individual flakes, and change over time, all of which makes control of the optoelectronic properties challenging. There are many different perturbations that can alter the optical properties, including charge doping, defects, strain, oxidation, adsorbed molecules and water intercalation. Identifying which perturbations are present is usually not straightforward and requires multiple measurements using multiple experimental modalities, which presents barriers when attempting to optimise preparation of these materials. Here, we apply high-resolution photoluminescence and differential reflectance hyperspectral imaging in situ to CVD-grown WS2 monolayers. By combining these two optical measurements and using a statistical correlation analysis we are able to disentangle three contributions modulating optoelectronic properties of these materials: electron doping, strain and disorder arising from adsorbates, substrate fluctuations and/or defects. In separating these contributions, we also observe that the B-exciton energy is less sensitive to variations in doping density than A-excitons. Our approach is not limited to monolayers of WS2 and is also applicable to other 2D materials with optical transitions, and the general approach could be extended to other characterization modalities.

Beyond the visible: The Viterbo Crucifixion panel painting attributed to Michelangelo Buonarroti

The painting object of the present work, currently exposed in the Museum of Colle del Duomo in Viterbo (Italy), has been dated back by art historians to the 16th century and it owes its relevance to a still discussed attribution to Michelangelo Buonarroti. For this reason, art historians and the responsible curator of the Museum commissioned scientific investigations to support and explain their hypothesis about the painting attribution and dating. Here we report the results of two sets of investigation: diffuse reflectance hyperspectral imaging and X-ray fluorescence spectroscopy of the painting; radiocarbon dating and identification of the panel wood. The hyperspectral dataset, coupled with X-ray fluorescence spectra on selected analysis points, reveals the presence of precious ultramarine blue and vermillion pigments, confirming the importance of the painting committer. Wood analysis and radiocarbon dating by wiggle matching technique revealed that the botanical species used for the panel is Cypress (Cupressus sempervirens L.), and enabled dating the painting around AD 1500± 25, providing to art historians a further element for the attribution of the artwork.

Apple Bruise Grading Using Piecewise Nonlinear Curve Fitting for Hyperspectral Imaging Data

Apple fruits can be easily damaged, and bruises occur on peels during harvest, transportation and storage, which could decrease fruit quality. This paper proposed an apple bruise grading method based on hyperspectral imaging (HSI). The spectral information of sound apples (Yantai Fuji 8) was first captured using a hyperspectral reflectance imaging device (386-1016 nm). These apples were then mechanically damaged by the same impact forces, and the bruised regions were exposed to room temperature for at most 120 min. The spectral data of the bruised apples at four different exposure times (30 min, 60 min, 90 min and 120 min) were obtained. The spectral data were preprocessed using Procrustes analysis (PA) to enable a more diverse distribution of the spectra among different patterns. To both accurately maintain the spectral information of different patterns and reduce the dimensions of the spectra, piecewise nonlinear curve fitting (PWCF) was presented using the least squares algorithm, where the resultant fitting coefficients from different spectral intervals were catenated into a low-dimension feature descriptor. The feature descriptors were then fed to an error-correction output coding-based support vector machine (ECOC-SVM) to grade the bruised apples. To further evaluate the performance of the presented PWCF, conventional algorithms, including the successive projections algorithm (SPA), genetic algorithm (GA), principal component analysis (PCA) and kernel principal component analysis (KPCA), were introduced for comparison. Experimental results showed that the proposed method obtained the best grading accuracy (97.33%) compared to the other methods.

Setting up a methodology to distinguish between green oranges and leaves using hyperspectral imaging

The estimation of green citrus fruit yield is a key parameter for growers and the industry. The early estimation of orange yield at the immature green stage could influence the future market price and allow producers to plan the harvest in advance, thus reducing costs. This research can be considered as a preliminary step for designing low-cost spectral cameras capable of being mounted on unmanned aerial vehicles (UAVs) to estimate orange yield and defects. Images were acquired from oranges and leaves from an orchard in Jeju Island (Jeju, Republic of Korea), using two hyperspectral reflectance imaging systems, one working in the range 400–1000 nm (visible/near infrared, Vis/NIR) and the other between 900 and 2500 nm (short-wave infrared, SWIR). The main objective of the research was to set up a methodology to select the relevant bands - from the two spectral ranges studied - to distinguish between green oranges and leaves and to detect defects, which will allow citrus yield to be estimated. Analysis of variance (ANOVA) and principal component analysis (PCA) were used to select the key wavelengths for this purpose; next, a band ratio coupled with a simple thresholding method was applied. This study showed that the Vis/NIR hyperspectral imaging correctly classified 96.97% and 92.93% of the pixels, respectively, to distinguish between green oranges and leaves and to detect defects, while with the SWIR system, the percentage of pixels correctly classified for these two objectives were 74.79% and 89.31%, respectively. These results confirm that it is possible to use a low number of wavelengths to estimate harvest yield in oranges, which could pave the way for the future development of low-cost and low-weight equipment for the detection of green and sound fruit.

Toward assessment of resection margins using hyperspectral diffuse reflection imaging (400-1,700 nm) during tongue cancer surgery.

There is a clinical need to assess the resection margins of tongue cancer specimens, intraoperatively. In the current ex vivo study, we evaluated the feasibility of hyperspectral diffuse reflectance imaging (HSI) for distinguishing tumor from the healthy tongue tissue. Fresh surgical specimens (n = 14) of squamous cell carcinoma of the tongue were scanned with two hyperspectral cameras that cover the visible and near-infrared spectrum (400-1,700 nm). Each pixel of the hyperspectral image represents a measure of the diffuse optical reflectance. A neural network was used for tissue-type prediction of the hyperspectral images of the visual and near-infrared data sets separately as well as both data sets combined. HSI was able to distinguish tumor from muscle with a good accuracy. The diagnostic performance of both wavelength ranges (sensitivity/specificity of visual and near-infrared were 84%/80% and 77%/77%, respectively) appears to be comparable and there is no additional benefit of combining the two wavelength ranges (sensitivity and specificity were 83%/76%). HSI has a strong potential for intra-operative assessment of tumor resection margins of squamous cell carcinoma of the tongue. This may optimize surgery, as the entire resection surface can be scanned in a single run and the results can be readily available. Lasers Surg. Med. © 2019 Wiley Periodicals, Inc.