Imaging Of Human Skin Research Articles

Dataset of human skin and fingernails images for non-invasive haemoglobin level assessment

Anaemia, a decrease in total concentration of haemoglobin (Hb) in blood, affects substantial percentage of the population worldwide. Currently, the gold standard for determining the Hb level is the invasive analysis of venous blood. Yet, more and more research groups demonstrate the possibility of non-invasive Hb assessment using white light imaging of tissue sites where Hb is the main chromophore, in particular, fingernails. Despite the promising declarations, non-invasive Hb assessment via RGB-imaging is still poorly used in practice. The main reason is the difficulty in establishing the true accuracy of the methods presented in different works since they are tested on private datasets collected under different experimental conditions. Here we present an open dataset containing RGB images of skin and fingernails for patients with a known level of Hb, thus providing a single benchmark for researchers and engineers in the field, aimed at fostering translation of non-invasive imaging methods to the bedside.

Hyperspectral imaging of 4T1 mammary carcinomas grown in dorsal skinfold window chambers: spectral renormalization and fluorescence modeling.

Hyperspectral imaging (HSI) of murine tumor models grown in dorsal skinfold window chambers (DSWCs) offers invaluable insight into the tumor microenvironment. However, light loss in a glass coverslip is often overlooked, and particular tissue characteristics are improperly modeled, leading to errors in tissue properties extracted from hyperspectral images. We highlight the significance of spectral renormalization in HSI of DSWC models and demonstrate the benefit of incorporating enhanced green fluorescent protein (EGFP) excitation and emission in the skin tissue model for tumors expressing genes to produce EGFP. We employed an HSI system for intravital imaging of mice with 4T1 mammary carcinoma in a DSWC over 14 days. We performed spectral renormalization of hyperspectral images based on the measured reflectance spectra of glass coverslips and utilized an inverse adding-doubling (IAD) algorithm with a two-layer murine skin model, to extract tissue parameters, such as total hemoglobin concentration and tissue oxygenation ( ). The model was upgraded to consider EGFP fluorescence excitation and emission. Moreover, we conducted additional experiments involving tissue phantoms, human forearm skin imaging, and numerical simulations. Hyperspectral image renormalization and the addition of EGFP fluorescence in the murine skin model reduced the mean absolute percentage errors (MAPEs) of fitted and measured spectra by up to 10% in tissue phantoms, 0.55% to 1.5% in the human forearm experiment and numerical simulations, and up to 0.7% in 4T1 tumors. Similarly, the MAPEs for tissue parameters extracted by IAD were reduced by up to 3% in human forearms and numerical simulations. For some parameters, statistically significant differences ( ) were observed in 4T1 tumors. Ultimately, we have shown that fluorescence emission could be helpful for 4T1 tumor segmentation. The results contribute to improving intravital monitoring of DWSC models using HSI and pave the way for more accurate and precise quantitative imaging.

Large‐Scale Lensless Microscopy with Fast Acquisition and Region‐Wise Focusing

AbstractLensless microscopy provides a wide field of view (FOV) determined by the image sensor size, allowing visualization of large sample areas. Coupled with advanced and even pixel super‐resolution phase retrieval algorithms, it can achieve resolutions up to the sub‐micron level, enabling both large‐FOV and high‐resolution imaging. However, high‐throughput lensless imaging encounters challenges in rapid data acquisition and large‐scale phase retrieval. Furthermore, when examining biological samples over a large FOV, focal plane inconsistencies often emerge among distinct regions. This study introduces a fast acquisition and efficient reconstruction method for coherent lensless imaging. Multiple measurements are manually modulated using an axial translation stage and sequentially captured by an image sensor, requiring no hardware synchronization. Optical parameter calibration, region‐wise auto‐focusing, and region‐wise phase retrieval algorithms are integrated to establish a general parallel computing framework for rapid, efficient, and high‐throughput lensless imaging. Experimental results demonstrate a 7.4 mm × 5.5 mm FOV and 1.38 µm half‐pitch resolution imaging of human skin and lung tumor sections with region‐wise focusing, requiring ≈0.5‐s acquisition time and 17‐s reconstruction time. By incorporating pixel super‐resolution, a 0.98 µm half‐pitch resolution is achieved in full‐FOV peripheral blood smears without additional data required, advantageous for discerning hollow shapes and segmenting blood cells.

Distance-based integration method for human skin type identification

Although identifying human skin types is essential in dermatology, cosmetology, and facial recognition, the classification of human skin types is challenging due to the complex nature, varied characteristics, and the influence of external factors. Traditional methods for skin type identification often rely on subjective assessments, leading to inconsistent and inaccurate results. Therefore, this paper proposes a novel method named a distance-based integration method to identify skin types based on the Fitzpatrick skin scale, also known as the Fitzpatrick skin type. This study focuses on the objective distance measurement, integrated with the Fuzzy Analytic Hierarchy Process (AHP). The objective distance was utilized to determine the distance between each HEX color code for a clinical image and each target skin type. The Fuzzy AHP algorithm was employed to calculate the total score for each target class to identify human skin type. For this study, 1,022 images of human skin were used in the experiment. The results indicated that the proposed method achieved a high average accuracy of 93 %, precision of 80 %, and specificity of 96 %.

Training With Uncertain Annotations for Semantic Segmentation of Basal Cell Carcinoma From Full-Field OCT Images.

Semantic segmentation of basal cell carcinoma (BCC) from full-field optical coherence tomography (FF-OCT) images of human skin has received considerable attention in medical imaging. However, it is challenging for dermatopathologists to annotate the training data due to OCT's lack of color specificity. Very often, they are uncertain about the correctness of the annotations they made. In practice, annotations fraught with uncertainty profoundly impact the effectiveness of model training and hence the performance of BCC segmentation. To address this issue, we propose an approach to model training with uncertain annotations. The proposed approach includes a data selection strategy to mitigate the uncertainty of training data, a class expansion to consider sebaceous gland and hair follicle as additional classes to enhance the performance of BCC segmentation, and a self-supervised pre-training procedure to improve the initial weights of the segmentation model parameters. Furthermore, we develop three post-processing techniques to reduce the impact of speckle noise and image discontinuities on BCC segmentation. The mean Dice score of BCC of our model reaches 0.503±0.003, which, to the best of our knowledge, is the best performance to date for semantic segmentation of BCC from FF-OCT images.

Photoacoustic Imaging of Human Skin for Accurate Diagnosis and Treatment Guidance

Photoacoustic imaging (PAI) is a cutting-edge biomedical imaging modality, providing detailed anatomical and functional information about the area beneath the skin surface. Its light energy deposition is such that PAI typically provides clear images of the skin with high signal-to-noise ratios. Specifically, the rich optical contrast of PAI allows biological information related to lesion growth, malignancy, treatment response, and prognosis to be seen. Given its significant advantages and emerging role in imaging skin lesions, we summarize and comment on representative studies of skin PAI, such as the guidance of skin cancer biopsies and surgical excisions, and the accurate diagnosis of psoriasis. We conclude with our insights about the clinical significance of skin PAI, showing how its use to identify biological characteristics in lesion microenvironments allows early diagnosis and prognosis of disease.

Terahertz in vivo imaging of human skin: Toward detection of abnormal skin pathologies.

Terahertz (THz) imaging has long held promise for skin cancer detection but has been hampered by the lack of practical technological implementation. In this article, we introduce a technique for discriminating several skin pathologies using a coherent THz confocal system based on a THz quantum cascade laser. High resolution in vivo THz images (with diffraction limited to the order of 100 μm) of several different lesion types were acquired and compared against one another using the amplitude and phase values. Our system successfully separated pathologies using a combination of phase and amplitude information and their respective surface textures. The large scan field (50 × 40 mm) of the system allows macroscopic visualization of several skin lesions in a single frame. Utilizing THz imaging for dermatological assessment of skin lesions offers substantial additional diagnostic value for clinicians. THz images contain information complementary to the information contained in the conventional digital images.

High-Frequency Linear Array (20 MHz) Based on Lead-Free BCTZ Crystal.

Centimeter-sized BaTiO3-based crystals grown by top-seeded solution growth from the BaTiO3-CaTiO3-BaZrO3 system were used to process a high-frequency (HF) lead-free linear array. Piezoelectric plates with (110)pc cut within 1° accuracy were used to manufacture two 1-3 piezo-composites with thicknesses of 270 and [Formula: see text] for resonant frequencies in air of 10 and 30 MHz, respectively. The electromechanical characterization of the BCTZ crystal plates and the 10-MHz piezocomposite yielded the thickness coupling factors of 40% and 50%, respectively. We quantified the electromechanical performance of the second piezocomposite (30 MHz) according to the reduction in the pillar sizes during the fabrication process. The dimensions of the piezocomposite at 30 MHz were sufficient for a 128-element array with a 70- [Formula: see text] element pitch and a 1.5-mm elevation aperture. The transducer stack (backing, matching layers, lens, and electrical components) was tuned with the characteristics of the lead-free materials to deliver optimal bandwidth and sensitivity. The probe was connected to a real-time HF 128-channel echographic system for acoustic characterization (electroacoustic response and radiation pattern) and to acquire high-resolution in vivo images of human skin. The center frequency of the experimental probe was 20 MHz, and the fractional bandwidth at -6 dB was 41%. Skin images were compared against those obtained with a lead-based 20-MHz commercial imaging probe. Despite significant differences in sensitivity between elements, in vivo images obtained with a BCTZ-based probe convincingly demonstrated the potential of integrating this piezoelectric material in an imaging probe.

A Miniaturized Electrothermal-MEMS-Based Optical Coherence Tomography (OCT) Handheld Microscope

Swept-source optical coherence tomography (SS-OCT), benefiting from its high sensitivity, relatively large penetration depth, and non-contact and non-invasive imaging capability, is ideal for human skin imaging. However, limited by the size and performance of the reported optical galvanometer scanners, existing portable/handheld OCT probes are still bulky, which makes continuously handheld imaging difficult. Here, we reported a miniaturized electrothermal-MEMS-based SS-OCT microscope that only weighs about 25 g and has a cylinder with a diameter of 15 mm and a length of 40 mm. This MEMS-based handheld imaging probe can achieve a lateral resolution of 25 μm, a 3D imaging time of 5 s, a penetration depth of up to 3.3 mm, and an effective imaging field of view (FOV) of 3 × 3 mm2. We have carried out both calibration plate and biological tissue imaging experiments to test the imaging performance of this microscope. OCT imaging of leaves, dragonfly, and human skin has been successfully obtained, showing the imaging performance and potential applications of this probe on human skin in the future.

Detection of Monkeypox Disease from Human Skin Images with a Hybrid Deep Learning Model.

Monkeypox, a virus transmitted from animals to humans, is a DNA virus with two distinct genetic lineages in central and eastern Africa. In addition to zootonic transmission through direct contact with the body fluids and blood of infected animals, monkeypox can also be transmitted from person to person through skin lesions and respiratory secretions of an infected person. Various lesions occur on the skin of infected individuals. This study has developed a hybrid artificial intelligence system to detect monkeypox in skin images. An open source image dataset was used for skin images. This dataset has a multi-class structure consisting of chickenpox, measles, monkeypox and normal classes. The data distribution of the classes in the original dataset is unbalanced. Various data augmentation and data preprocessing operations were applied to overcome this imbalance. After these operations, CSPDarkNet, InceptionV4, MnasNet, MobileNetV3, RepVGG, SE-ResNet and Xception, which are state-of-the-art deep learning models, were used for monkeypox detection. In order to improve the classification results obtained in these models, a unique hybrid deep learning model specific to this study was created by using the two highest-performing deep learning models and the long short-term memory (LSTM) model together. In this hybrid artificial intelligence system developed and proposed for monkeypox detection, test accuracy was 87% and Cohen's kappa score was 0.8222.

Portable PIMI polarization imaging device based on automatic polarization recognition.

This paper proposes a new portable polarization parametric indirect microscopy imaging without a liquid crystal (LC) retarder. The polarization was modulated by a polarizer automatically rotating when the camera took raw images sequentially. A specific mark tagged the polarization states of each camera's snapshot in the optical illumination path. A computer vision portable polarization parametric indirect microscopy imagrecognition algorithm was developed to retrieve the unknown polarization states from each raw camera image to ensure that the right polarization modulation states were used in the PIMI processing algorithm. The system's performance was verified by obtaining PIMI parametric images of human facial skin. The proposed method avoids the error problem caused by the LC modulator and significantly reduces the whole system's cost.

Direct frequency domain fluorescence lifetime imaging using simultaneous ultraviolet and visible excitation.

Due to the complexity, limited practicality, and cost of conventional fluorescence lifetime imaging/microscopy (FLIM) instrumentation, FLIM adoption has been mostly limited to academic settings. We present a novel point scanning frequency-domain (FD) FLIM instrumentation design capable of simultaneous multi-wavelength excitation, simultaneous multispectral detection, and sub-nanosecond to nanosecond fluorescence lifetime estimation. Fluorescence excitation is implemented using intensity-modulated CW diode lasers that are available in a selection of wavelengths spanning the UV-VI-NIR range (375-1064 nm). Digital laser intensity modulation was adopted to enable simultaneous frequency interrogation at the fundamental frequency and corresponding harmonics. Time-resolved fluorescence detection is implemented using low-cost, fixed-gain, narrow bandwidth (100 MHz) avalanche photodiodes, thus, enabling cost-effective fluorescence lifetime measurements at multiple emission spectral bands simultaneously. Synchronized laser modulation and fluorescence signal digitization (250 MHz) is implemented using a common field-programmable gate array (FPGA). This synchronization reduces temporal jitter, which simplifies instrumentation, system calibration, and data processing. The FPGA also allows for the implementation of the real-time processing of the fluorescence emission phase and modulation at up to 13 modulation frequencies (processing rate matching the sampling rate of 250 MHz). Rigorous validation experiments have demonstrated the capabilities of this novel FD-FLIM implementation to accurately measure fluorescence lifetimes in the range of 0.5-12 ns. In vivo endogenous, dual-excitation (375nm/445nm), multispectral (four bands) FD-FLIM imaging of human skin and oral mucosa at 12.5 kHz pixel rate and room-light conditions was also successfully demonstrated. This versatile, simple, compact, and cost-effective FD-FLIM implementation will facilitate the clinical translation of FLIM imaging and microscopy.

High-speed reflectance confocal microscopy of human skin at 1251-1342 nm.

Reflectance confocal microscopy (RCM) is an imaging method that can noninvasively visualize microscopic features of the human skin. The utility of RCM can be further improved by increasing imaging speed. In this paper, we report high-speed RCM imaging of human skin with a frame rate that is over 10 times faster and an area imaging rate that is 6-9 times faster than those of commercially available RCM devices. The higher imaging speed was achieved using a high-speed RCM technique, termed spectrally encoded confocal microscopy (SECM). SECM uses a diffraction grating and a high-speed, wavelength-swept source to conduct confocal imaging at a very high rate. We developed a handheld SECM probe using a scanned-grating approach. The SECM probe was used in conjunction with a wavelength-swept source with a spectral band of 1251-1342 nm. The SECM probe achieved high lateral resolution of 1.3-1.6 µm and an axial resolution of 3.5 µm. SECM images of the human skin (image size = 439 × 439 µm2 ) obtained at 100 frames/s clearly show previously reported RCM features of the human skin in vivo with adequate image quality. The fast imaging speed allowed for the rapid acquisiton of volumetric SECM image data (200 frames covering a depth range of 200 µm) within 2 s. The use of 1251-1342 nm provided sufficient signal level and contrast required to visualize key cellular morphologic features. These preliminary results demonstrate that high-speed SECM imaging of the human skin at 1251-1342 nm is feasible.

Depth perception between dots and the background face reduces trypophobic discomfort

BackgroundStudies have shown that viewing a cluster of dots evokes feelings of discomfort in viewers and that the discomfort becomes especially strong when the dots are placed on background images of human skin. This phenomenon has been explained by the physical properties and spatial and semantic relationships between the dots and the background. However, it was not known whether the perceived, as well as the physical, spatial relationships contributes to the generation of discomfort.MethodsWe evoked illusory depth perception between black dots and the background face by drawing shadow-like gray dots around the black dots, while keeping the same black dots at the same positions, and examined whether illusory depth perception could increase or decrease discomfort. In each trial, participants viewed one of the following types of facial images: (a) face-only (face without dots), (b) a cluster of black dots on the face, (c) a cluster of gray dots on the face, and (d) a cluster of black dots and shadow-like gray dots on the face. After seeing each picture, they evaluated how much discomfort they felt from viewing the picture using a Likert scale and reported whether they perceived depth between the dots and the face.ResultsParticipants felt discomfort toward all three types of faces with dots, that is, faces with black dots, gray dots, and both. However, interestingly, participants felt less discomfort when both black and gray dots were presented on the face than when only black dots were presented. The participants perceived depth between the black dots and the face in 85% of the trials with black dots and shadow-like gray dots, and there was a significant correlation between discomfort and frequency of depth perception. However, in the trials with black dots only and gray dots only, they perceived depth in only 18% and 27% of the trials, respectively, and the correlations between the frequencies of depth perception and discomfort were not significant.ConclusionsOur results suggest that the perceived spatial relationship, such as attached vs. separate, as well as the physical spatial relationship, contribute to the generation of discomfort.

H&E-like staining of OCT images of human skin via generative adversarial network

Noninvasive and high-speed optical coherence tomography (OCT) systems have been widely deployed for daily clinical uses. High-resolution OCTs are advancing rapidly; however, grey-level OCT images are not easy to read for pathologists due to the lack of diagnosis specificity compared with hematoxylin and eosin (H&E) stained images. This work presents an OCT to H&E image translation model to convert the OCT images to H&E-like stained images using unpaired OCT and H&E datasets. “H&E like” means the stratum corneum (SC) boundary and the dermal-epidermal junction (DEJ) of the OCT and the translated images are consistent. Pre-trained segmentation models for the DEJ and the SC are exploited to enhance the performance of anatomical image translation and reduce the DEJ and SC lower boundary errors to ±2.3 and ±1.7 μm, respectively. A pre-trained VGG16 network extracts the features of the nuclei. Pearson's correlation coefficient of the nuclei location and size consistency is 84% ± 1%. As a result, in vivo medical image translation accuracy with cellular resolution was achieved.

Mirau-based line-field confocal optical coherence tomography for three-dimensional high-resolution skin imaging.

SignificanceLine-field confocal optical coherence tomography (LC-OCT) is a recently introduced high-resolution imaging modality based on a combination of low-coherence optical interferometry and reflectance confocal optical microscopy with line illumination and line detection. Capable of producing three-dimensional (3D) images of the skin with cellular resolution, in vivo, LC-OCT has been mainly applied in dermatology and dermo-cosmetology. The LC-OCT devices capable of acquiring 3D images reported so far are based on a Linnik interferometer using two identical microscope objectives. In this configuration, LC-OCT cannot be designed to be a very compact and light device, and the image acquisition speed is limited.AimThe objective of this work was to develop a more compact and lighter LC-OCT device that is capable of acquiring images faster without significant degradation of the resolution and with optimized detection sensitivity.ApproachWe developed an LC-OCT device based on a Mirau interferometer using a single objective. Dynamic adjustment of the camera frequency during the depth scan is implemented, using a faster camera and a more powerful light source. The reflectivity of the beam-splitter in the Mirau interferometer was optimized to maximize the detection sensitivity. A galvanometer scanner was incorporated into the device for scanning the illumination line laterally. A stack of adjacent B-scans, constituting a 3D image, can thus be acquired.ResultsThe device is able to acquire and display B-scans at 17 fps. 3D images with a quasi-isotropic resolution of (1.3, 1.9, and in the , and directions, respectively) over a field of () can be obtained. 3D imaging of human skin at cellular resolution, in vivo, is reported.ConclusionsThe acquisition rate of the B-scans, at 17 fps, is unprecedented in LC-OCT. Compared with the conventional LC-OCT devices based on a Linnik interferometer, the reported Mirau-based LC-OCT device can acquire B-scans times faster. With potential advantages in terms of compactness and weight, a Mirau-based device could easily be integrated into a smaller and lighter handheld probe for use by dermatologists in their daily medical practice.

Comparative study of OCTA algorithms with a high-sensitivity multi-contrast Jones matrix OCT system for human skin imaging.

The multi-contrast Jones matrix OCT (JMT) system can not only improve the tissue-specific contrast but also enhance the sensitivity of detecting flow, which is gaining increasing attention. However, for the JMT system, there is currently a lack of studies evaluating and guiding the selection of suitable angiography algorithms to map the most appealing quality of angiograms for clinical use. In this paper, by a homemade high-sensitivity multi-contrast JMT system based 200 kHz swept source, the performance of two complex-signal-based OCTA methods that are insensitive to phase instability and one amplitude-signal-based OCTA method are compared for in-vivo imaging of human skin qualitatively and quantitatively. Six metrics, including vascular connectivity, image contrast-to-noise ratio, image signal-to-noise ratio, vessel diameter index, blood vessel density, and processing time, are assessed. The results show that the vascular networks processed by all OCTA methods and the texture of skin could be visualized simultaneously and markedly. Additionally, the complex-signal-based OCTA methods successfully suppress phase instabilities and even outperform the amplitude-signal-based OCTA algorithm in some indicators. This paper has a certain guiding significance for selecting an appropriate angiography algorithm and expanding the application field with this system.

HS‐UNET‐ID: An approach for human skin classification integrating between UNET and improved dense convolutional network

AbstractThe Sun's ultraviolet radiation, toxic chemicals from industry, and other factors can cause problems to human skin. These factors have boosted the risk of humans contracting skin cancer. In identifying and detecting signals of skin cancer diseases, recent advancements in deep learning techniques significantly improve the accuracy of diagnosis. This work delves into processing techniques for human skin images with MorphologyEx (blackhat), thresholding, semantic segmentation with UNET, and several famous convolutional neural network architectures to improve skin classification performance. If skin cancer can be stopped in the first stage, the survival rate of the patient could be more than 95%, while it drops to just 5% if the detection is carried lately in Stage IV4. Due to its seriousness, the healthcare and research community has received considerable attention, and its most important goal is to be able to diagnose it at an early stage. However, the challenge due to the similarity in melanocytic and non‐melanoma skin lesions is inevitable. We achieved skin classification performances of 0.8560, 0.9672, 0.8063, and 0.8560 measured by overall accuracy, area under curve, Matthews correlation coefficient, and F1‐score, respectively. The proposed approach has been deployed into applications on Mobile and Web platforms.

Compressive-sensing swept-source optical coherence tomography angiography with reduced noise.

Optical coherence tomography angiography (OCTA), as a functional extension of optical coherence tomography (OCT), has exhibited a great potential to aid in clinical diagnostics. Currently, OCTA still suffers from motion artifact and noise. Therefore, in this article, we propose to implement compressive sensing (CS) on B-scans to reduce motion artifact by increasing B-scan rate. Meanwhile, a noise reduction filter is specially designed by combining CS, Gaussian filter and median filter. Specially, CS filtering is realized by averaging multiple CS repetitions on en-face OCTA images with varied sampling functions. The method is evaluated on in vivo OCTA images of human skin. The results show that vasculature structures can be reconstructed well through CS on B-scans with a sampling rate of 70%. Moreover, the noise can be significantly eliminated by the developed filter. This implies that our method has a good potential to expedite OCTA imaging and improve the image quality.

Inside Cover

Here, a three dimensional confocal photoacoustic dermoscopy (3D‐CPAD) equipped with an autofocusing sono‐opto probe was developed to facilitate the 3D reconstruction of high‐contrast and high‐spatial‐resolution imaging of human skin with multilaminate structures in depth direction. The autofocusing sono‐opto probe mainly integrated an electrowetting‐based varifocal lens to automatically control the acoustic and optical confocal length, and an annular ultrasonic detector with coaxial configuration for receiving photoacoustic signals.Further details can be found in the article by Haigang Ma, Zhiyang Wang, Chao Zuo, and Qinghua Huang (e202100323) image