Multispectral and autofluorescence RGB imaging for skin cancer diagnostics
This paper presents the results of statistical clinical data, combining two diagnostic methods. A combination of two skin imaging methods – diffuse reflectance and autofluorescence – has been applied for skin cancer diagnostics. Autofluorescence (AF) and multispectral diffuse reflectance images were acquired by custom made prototype with 405 nm, 526 nm, 663 nm and 964 nm LEDs and RGB CMOS camera. Parameter p’ was calculated from diffuse reflectance images under green, red and infrared illumination, AF intensity (I’) was calculated from AF images exited at 405nm wavelength. Obtained results show that criterion p` > 1 gives possibility to discriminate melanomas and different kind of keratosis…
Challenges of automatic processing of large amount of skin lesion multispectral data
This work will describe the challenges involved in setting up automatic processing for a large differentiated data set. In this study, a multispectral (skin diffuse reflection images using 526nm (green), 663nm (red), and 964nm (infrared) illumination and autofluorescence (AF) image using 405 nm excitation) data set with 756 lesions (3024 images) was processed. Previously, using MATLAB software, finding markers, correctly segmenting images with dark edges and image alignment were the main causes of the problems in automatic data processing. To improve automatic processing and eliminate the use of licensed software, the latter was substituted with the open source Python environment. For more …
Image quality enhancement for skin cancer optical diagnostics
The research presents image quality analysis and enhancement proposals in biophotonic area. The sources of image problems are reviewed and analyzed. The problems with most impact in biophotonic area are analyzed in terms of specific biophotonic task – skin cancer diagnostics. The results point out that main problem for skin cancer analysis is the skin illumination problems. Since it is often not possible to prevent illumination problems, the paper proposes image post processing algorithm – low frequency filtering. Practical results show diagnostic results improvement after using proposed filter. Along that, filter do not reduces diagnostic results’ quality for images without illumination de…
Identification of the most informative wavelengths for non-invasive melanoma diagnostics in spectral region from 450 to 950 nm
In this study 300 skin lesion (including 32 skin melanomas) multispectral data cubes were analyzed. The multi-step and single step machine learning approaches were analyzed to find the wavebands that provide the most information that helps discriminate skin melanoma from other benign pigmented lesions. The multi-step machine learning approach assumed training several models but proved itself to be ineffective. The reason for that is a necessity to train a segmentation model on a very small dataset and utilization of standard machine learning classifier which have shown poor classification performance. The single-step approach is based on a deep learning neural network. We have conducted 260…
Benign — A typical nevi discrimination using diffuse reflectance and fluorescence multispectral imaging system
Early diagnostics of skin cancer is of interest for dermatologists. Atypical nevi are not considered to be malignant, but are suspects that should be detected and monitored over time. The multispectral imaging system Nuance operating in spectral range 450–950 nm was adapted for clinical in vivo measurements in diffuse reflectance and fluorescence mode. Mean and standard deviation values of optical density and fluorescence intensity were extracted from segmented pigmented lesions (21 benign and 26 atypical nevi) and used for further analysis. It was possible to achieve 62% sensitivity and 67% specificity for discrimination between atypical and benign lesions using averaged fluorescence mean …
Quality enhancement of multispectral images for skin cancer optical diagnostics
Melanoma is the least common but deadliest skin cancer, accounting for only about 1% of all cases, but is the cause of the vast majority of skin cancer death. In some parts of the world, especially among western countries, melanoma is becoming more common every year. The detection of melanoma in early stage can be helpful to cure it. Unfortunately, long ques and high prices for dermatology service can result in the skin cancer diagnosis at its later stage, thus increasing the risk of mortality for the patient. It is important to provide a non-invasive optical device for primary care physicians to help diagnose different skin malformation based on obtained optical images. Such device will be…
Optical design improvement for noncontact skin cancer diagnostic device
Multispectral diffuse reflectance imaging and autofluorescence photo-bleaching imaging are methods that have been investigated for use in skin disorder diagnostics. In response to the ever-increasing incidence of skin cancer in light skinned populations a new device has been designed incorporating both of these methods. The aim of the study was to create a device that is most efficient in terms of hardware and software parameters for the screening of malignant and benign skin lesions. A set of 525 nm, 630 nm and 980 nm LEDs were used to illuminate the skin area at three wavelengths [1] and a set of 405 nm LEDs were used to induce the skin autofluorescence [2]. For a more homogenous illumina…
A method for skin malformation classification by combining multispectral and skin autofluorescence imaging
As the incidence of skin cancer is still increasing worldwide, there is a high demand for early, non-invasive and inexpensive skin lesion diagnostics. In this article we describe and combine two skin imaging methods: skin autofluorescence (AF) and multispectral criterion p’. To develop this method, we used custom made prototype with 405 nm, 526 nm, 663 nm and 964 nm LED illuminations, perpendicular positioned linear polarizers, 515 nm filter and IDS camera. Our aim is to develop a skin lesion diagnostic device for primary care physicians who do not have experience in dermatology or skin oncology. In this study we included such common benign lesion groups as seborrheic keratosis, hyperkerato…
Use of machine learning approaches to improve non-invasive skin melanoma diagnostic method in spectral range 450 - 950nm
Non-invasive skin cancer diagnostic methods develop rapidly thanks to Deep Learning and Convolutional Neural Networks (CNN). Currently, two types of diagnostics are popular: (a) using single image taken under white illumination and (b) using multiple images taken in narrow spectral bands. The first method is easier to implement, but it is limited in accuracy. The second method is more sensitive, because it is possible to use illumination considering the absorption bands of the skin chromophores and the optical properties of the skin. Currently CNN use a single white light image, due to the availability of large datasets with lesion images. Since CNN processing and analysis requires a large …
Imaging of LED-excited autofluorescence photobleaching rates for skin diagnostics
The aim of this study is to develop a novel non-invasive approach for skin cancer (melanoma, basal cell and squamous cell carcinomas) diagnostics by mapping the AF intensity decrease (photo-bleaching) rates under continuous 405 nm LED excitation. For parametric mapping of skin AF intensity decrease rates a sequence of filtered AF imaging under 405 nm LED excitation for 20 seconds at a power density of ~7 mW/cm2 with a frame rate 0.5 fps was recorded and analyzed by cloud-based prototype device. Several clinical cases and potential future applications of the proposed autofluorescence photobleaching rate imaging technique are discussed.
Mobile platform for online processing of multimodal skin optical images: Using online Matlab server for processing remission, fluorescence and laser speckle images, obtained by using novel handheld device
Mobile platform for multimodal skin assessment has been developed. Different illumination sources allow switching between modalities. Diffuse reflectance spectral imaging is provided by five LEDs, fluorescence is excited by 405 nm LEDs, and laser speckle by 650 nm laser diode. Handheld, battery powered device includes all light sources and color camera with USB. The core of the system is Linux OS embedded microcontroller. USB, FTP and HTML with JavaScript combination is used to create standard image transfer and control interface. In combination with built in WiFi access point it allows online skin images storage and processing. It means that data processing algorithms are located and updat…
Laser speckle imaging for early detection of microbial colony forming units
In this study, an optical contactless laser speckle imaging technique for the early identification of bacterial colony-forming units was tested. The aim of this work is to compare the laser speckle imaging method for the early assessment of microbial activity with standard visual inspection under white light illumination. In presented research, the growth of Vibrio natriegens bacterial colonies on the solid medium was observed and analyzed. Both – visual examination under white light illumination and laser speckle correlation analysis were performed. Based on various experiments and comparisons with the theoretical Gompertz model, colony radius growth curves were obtained. It was shown that…
Towards to deep neural network application with limited training data: synthesis of melanoma's diffuse reflectance spectral images
The goal of our study is to train artificial neural networks (ANN) using multispectral images of melanoma. Since the number of multispectral images of melanomas is limited, we offer to synthesize them from multispectral images of benign skin lesions. We used the previously created melanoma diagnostic criterion p'. This criterion is calculated from multispectral images of skin lesions captured under 526nm, 663nm, and 964nm LED illumination. We synthesize these three images from multispectral images of nevus so that the p' map matches the melanoma criteria (the values in the lesion area is >1, respectively). Demonstrated results show that by transforming multispectral images of benign nevus i…
Automated microorganisms activity detection on the early growth stage using artificial neural networks
The paper proposes an approach of a novel non-contact optical technique for early evaluation of microbial activity. Noncontact evaluation will exploit laser speckle contrast imaging technique in combination with artificial neural network (ANN) based image processing. Microbial activity evaluation process will comprise acquisition of time variable laser speckle patterns in given sample, ANN based image processing and visualization of obtained results. The proposed technology will measure microbial activity (like growth speed) and implement these results for counting live microbes. It is expected, that proposed technology will help to evaluate number of colony forming units (CFU) and return r…
Laser speckle time-series correlation analysis for bacteria activity detection
The study aims at development and laboratory approbation of non-contact optical technique for early evaluation of microbial activity. Microorganisms’ activity is estimated by laser speckle contrast imaging technique in combination with image processing of obtained time varying speckle patterns. Laser speckle patterns were captured by CMOS sensor during illumination of growing bacteria colonies by low power (<30 mW, 635 nm) stabilized coherent light source. To validate proposed technique and image processing algorithm the vibrio natriegens bacteria are used. After analysis of several different experiments the following results were obtained: In the central part of the colony activity can be …
Autofluorescence imaging for recurrence detection in skin cancer postoperative scars
This clinical study is a first attempt to use autofluorescence for recurrence diagnosis of skin cancer in postoperative scars. The proposed diagnostic parameter is based on a reduction in scar autofluorescence, evaluated in the green spectral channel. The validity of the method has been tested on 110 postoperative scars from 56 patients suspected of non-melanoma skin cancer, with eight patients (13 scars) available for the repeated examination. The recurrence diagnosis within a scar has been made after two subsequent autofluorescence check-ups, representing the temporal difference between the scar autofluorescence amplitudes as a vector. The recognition of recurrence has been discussed to r…
Cloud Infrastructure for Skin Cancer Scalable Detection System
Skin cancer diagnostics is one of the medical areas where early diagnostic allows achieving patients’ high survival rate. Typically, skin cancer diagnostic is performed by dermatologist, since the amount of such specialists is limited, mortality rate is high [1]. By creating the low cost and easy to use diagnostic device, it is possible to bring skin cancer diagnostic to primary care physicians and allow to check much more persons and diagnose skin cancer on the early stages. There are several existing devices, that provide skin cancer diagnostics [2]. Most of them process the skin images locally and have limited diagnostic capabilities; some of them send images to dermatologists for manual…
Deep learning model deploying on embedded skin cancer diagnostic device
The number of research papers, where neural networks are applied in medical image analysis is growing. There is a proof that Convolutional Neural Networks (CNN) are able to differentiate skin cancer from nevi with greater accuracy than experienced specialists on average (sensitivity 82% and 73% accordingly).1 Team's latest research2 allows achieving even greater accuracy, by using specific narrow-band illumination. Nevertheless, the overall probability of early skin cancer detection depends on the availability of diagnostic tools. If screening tools will be available to a high number of general practices, the chance of disease detection will increase. The previous research3 shows that scala…
Embedded neural network system for microorganisms growth analysis
This study presents autonomous system for microorganisms’ growth analysis in laboratory environment. As shown in previous research, laser speckle analysis allows detecting submicron changes of substrate with growing bacteria. By using neural networks for speckle analysis, it is possible to develop autonomous system, that can evaluate microorganisms’ growth by using cheap optics and electronics elements. System includes embedded processing module, CMOS camera, 670nm laser diode and optionally WiFi module for connecting to external image storage system. Due to small size, system could be fully placed in laboratory incubator with constant humidity and temperature. By using laser diode, Petri d…
Multi-Path U-Net Architecture for Cell and Colony-Forming Unit Image Segmentation
U-Net is the most cited and widely-used deep learning model for biomedical image segmentation. In this paper, we propose a new enhanced version of a ubiquitous U-Net architecture, which improves upon the original one in terms of generalization capabilities, while addressing several immanent shortcomings, such as constrained resolution and non-resilient receptive fields of the main pathway. Our novel multi-path architecture introduces a notion of an individual receptive field pathway, which is merged with other pathways at the bottom-most layer by concatenation and subsequent application of Layer Normalization and Spatial Dropout, which can improve generalization performance for small datase…
Quantitative Multispectral Imaging Differentiates Melanoma from Seborrheic Keratosis.
Melanoma is a melanocytic tumor that is responsible for the most skin cancer-related deaths. By contrast, seborrheic keratosis (SK) is a very common benign lesion with a clinical picture that may resemble melanoma. We used a multispectral imaging device to distinguish these two entities, with the use of autofluorescence imaging with 405 nm and diffuse reflectance imaging with 525 and 660 narrow-band LED illumination. We analyzed intensity descriptors of the acquired images. These included ratios of intensity values of different channels, standard deviation and minimum/maximum values of intensity of the lesions. The pattern of the lesions was also assessed with the use of particle analysis. …