No abstract

Non-Melanoma skin cancer is one of the most frequent types of cancer. Early detection is encouraged so as to ensure the best treatment, Hyperspectral imaging is a promising technique for non-invasive inspection of skin lesions, however, the optimal wavelengths for these purposes are yet to be conclusively determined. A visible-near infrared hyperspectral camera with an ad-hoc built platform was used for image acquisition in the present study. Robust statistical techniques were used to conclude an optimal range between 573.45 and 779.88 nm to distinguish between healthy and non-healthy skin. Wavelengths between 429.16 and 520.17 nm were additionally found to be optimal for the differentiation between cancer types.

No abstract

Smartphones have widened the possibilities for low-cost close-range image acquisition for three-dimensional (3D) modelling. They allow the rapid acquisition of large amounts of data for a wide range of applications. However, the accuracy of the models and the automation possibilities depend on the image acquisition conditions and application requirements. In this study, the accuracy and reliability of the derived photogrammetric 3D models are evaluated on a spherical set-up for close-range applications (c.30 cm). Different numbers of images, network configurations, targets, devices and camera calibration methodologies are tested and evaluated. Results show that for this close-range application high accuracy (0Á2 mm) and reliability can be achieved. The number of images did not significantly affect the accuracy but was vital for tie-point detection and image orientation. The use of artificial targets was found to be the key factor in increasing the final accuracy. In contrast, the image calibration strategy and the characteristics of the imaging device did not have a great impact on the results.Smartphone video (more specifically slow-motion video) is a useful tool for acquiring large numbers of images, suitable even for fast moving objects. These images can be used for the creation of 3D models of moving objects (Barbero-Garc ıa et al., 2017). With an image acquisition speed of 240 frames per second (fps) of many smartphones (still far below the ultra-high-speed cameras that reach up to 2000 fps), the computational cost is the main limitation given the number of images to be used for 3D modelling.Despite their advantages, smartphone cameras present high internal instability that hampers their correct calibration. This problem is common to all non-metric digital cameras (Fraser, 2013), but is especially exacerbated when working with smartphones. The radiometric accuracy of smartphone cameras is lower than that of single-lens reflex (SLR) cameras but, despite their limitations, studies have concluded that these cameras can be used for photogrammetric tasks with a required accuracy of 1:10 000 (Akca and Gruen, 2009).The development of useful tools, which could allow non-expert users to obtain accurate 3D models for different purposes, requires a high degree of automation (Remondino et al., 2014). However, most of the available automatic low-cost solutions provide low repeatability and reliability (Remondino et al., 2012). The development of fully automatic and reliable solutions for specific applications requires an extensive knowledge of the factors affecting the quality of the 3D models created using smartphones or other similar imaging devices, such as tablets. The most important factors include the determination of the ideal geometric network, the selection of the best video frames and their optimal number, as well as the accuracy requirements for camera calibration. These parameters can vary greatly depending on the characteristics and limitations of the image acquisition process for a specific application (suc...

No abstract

No abstract

The use of smartphones cameras for photogrammetric purposes is increasing. However, the suitability of smartphones for 3D modelling for medical purposes in general, and for cranial deformation in particular, is still to be analysed. This paper investigates the suitability of smartphone video cameras to create 3D models for cranial deformation analysis compared to the digital single-lens reflex (SLR) cameras traditionally used in close-range photogrammetry. Two models are obtained, the first one from a slow-motion video recorded with a smartphone, and the second one from SLR camera imagery. The models are compared to evaluate the differences not only between themselves but also through the best fitting ellipsoid that allow the determination of the cranial deformations. The average distance between models is 0.5 mm, and below 1 mm for 86% of the model points. The maximum difference between the two fitted ellipsoid semiaxes is 1 mm. It can be stated that smartphones are a low-cost solution that can provide 3D models with a similar accuracy to that of SLR cameras for non-static objects in close range scenarios. More interestingly, slow-motion videos provide comparable results in real clinical conditions with infants in movement.

The early detection of Non-Melanoma Skin Cancer (NMSC) is crucial to achieve the best treatment outcomes. Shape is considered one of the main parameters taken for the detection of some types of skin cancer such as melanoma. For NMSC, the importance of shape as a visual detection parameter is not well-studied. A dataset of 993 standard camera images containing different types of NMSC and benign skin lesions was analysed. For each image, the lesion boundaries were extracted. After an alignment and scaling, Elliptic Fourier Analysis (EFA) coefficients were calculated for the boundary of each lesion. The asymmetry of lesions was also calculated. Then, multivariate statistics were employed for dimensionality reduction and finally computational learning classification was employed to evaluate the separability of the classes. The separation between malignant and benign samples was successful in most cases. The best-performing approach was the combination of EFA coefficients and asymmetry. The combination of EFA and asymmetry resulted in a balanced accuracy of 0.786 and an Area Under Curve of 0.735. The combination of EFA and asymmetry for lesion classification resulted in notable success rates when distinguishing between benign and malignant lesions. In light of these results, skin lesions’ shape should be integrated as a fundamental part of future detection techniques in clinical screening.