Recognition of images of finger skin with application of histogram, image filtration and K-NN classifier

doi:10.1016/j.bbe.2015.12.005

Biocybernetics and Biomedical Engineering

Volume 36, Issue 1, 2016, Pages 95-101

https://doi.org/10.1016/j.bbe.2015.12.005 Get rights and content

Abstract

In this paper, non-invasive method of recognition of finger skin was proposed. A plan of study of images of finger skin was proposed. Researches were carried out for three kinds of images: 60 h after injury, 160 h after injury, 450 h after injury. Proposed technique of recognition used methods of signal processing: extraction of magenta color, calculation of histogram, image filtration, calculation of perimeter, and K-NN classifier. A pattern creation process was conducted using 15 training images of finger skin. In the identification process 60 test images were used. The advantage of the presented method is analysis of the finger skin using a smartphone. The proposed approach will help to diagnose pathologies of human skin.

Introduction

Recognition of the medical images is a difficult problem in the analysis of digital images. The first problem is image acquisition. There are many devices to take images with various resolution. The second problem is to select proper processing methods to recognize pathology or biological variation. There are many methods of preprocessing, feature extraction and classification. There are also problems with different: age, gender and race of analyzed patients.

In the literature scientists develop many techniques of recognition of medical images [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12]. Many of them are related to perfusion images [7], [8], [9], [10]. Techniques based on thermal images are presented in recent years [11], [12]. Some of them are also related to recognition of pathology of human skins [13], [14], [15], [16], [17], [18]. Human skin can vary in color and appearance. Often this makes the task of recognition very difficult. Pathology of the human skin may lead to malignant melanoma. It is the most dangerous type of skin cancer [19], [20], [21]. In this paper Authors developed method of recognition of injured (Fig. 1) and healthy human finger (Fig. 2). The advantage of the presented method is analysis of the finger skin using a smartphone.

Most of previous approaches were related to finger vein [22], [23], [24]. Injured human fingers were common skin injury. It was a reason to analyze this kind of injury.

Section 1 describes general information about problems of recognition of human skin and short survey of literature. Section 2 presents the method of recognition of finger skin. Analysis of images of finger skin is discussed in Section 3. Conclusions of the article are presented in the last Section.

Section snippets

Proposed method of recognition of image of finger skin

Authors used smartphone (Colorovo CityTone) for acquisition of images. There is possibility to use other digital camera or smartphone. Next obtained images were copied to the computer (Intel Core i7-4702MQ, 16GB RAM, Windows 8, Matlab 2014b). Images were used as training and test images for process of recognition (proposed technique of recognition). The process of recognition of images of finger skin had 2 steps: pattern creation process and identification process (Fig. 3).

Both of them had

The results of recognition of images of injured human finger

Researches were carried out for three kinds of images mentioned in Section 2.1. The pattern creation process was conducted using 15 training images. The identification process used 60 test images.

Efficiency of image recognition was expressed as: $E = \frac{N o C I T I}{N o A T I} 100 %$ where E – efficiency of image recognition, NoCITI – number of correctly identified test images, NoATI – number of all test images.

The results of recognition of image of human finger with application of image filtration, histogram and

Discussion and conclusions

In this paper authors analyzed capabilities of recognition of proposed technique. This technique was based on image processing methods such as: extraction of magenta color, calculation of histogram, image filtration, calculation of perimeter, and K-NN classifier. The best results were obtained for k = 1 and k = 3 and image filtration, histogram and K-NN (100%). The pattern creation process was conducted using 15 training images. The identification process used 60 test images.

The first problem was

Acknowledgments

This work has been supported by AGH University of Science and Technology, grant nr 11.11.120.612, grant nr 11.11.120.354. We thank reviewers for their valuable suggestions.

References (29)

D. Reska et al.
MESA: complete approach for design and evaluation of segmentation methods using real and simulated tomographic images
Biocybern Biomed Eng
(2014)
T. Hachaj et al.
Application of neural networks in detection of abnormal brain perfusion regions
Neurocomputing
(2013)
R. Koprowski
Automatic analysis of the trunk thermal images from healthy subjects and patients with faulty posture
Comput Biol Med
(2015)
M. Marzec et al.
Methods of face localization in thermograms
Biocybern Biomed Eng
(2015)
L.K. Ferris et al.
Computer-aided classification of melanocytic lesions using dermoscopic images
J Am Acad Dermatol
(2015)
I. Maglogiannis et al.
Enhancing classification accuracy utilizing globules and dots features in digital dermoscopy
Comput Methods Progr Biomed
(2015)
A. Jukic et al.
Noninvasive diagnosis of melanoma with tensor decomposition-based feature extraction from clinical color image
Biomed Signal Process Control
(2013)
F.P. Ferrarese et al.
Performance evaluation in medical image segmentation
Curr Med Imaging Rev
(2013)
S. Masood et al.
A survey on medical image segmentation
Curr Med Imaging Rev
(2015)
A.A. Perez et al.
Segmentation and analysis of leg ulcers color image

L. Ogiela et al.

Cognitive computing in intelligent medical pattern recognition systems

Intelligent control and automation

(2006)

Z.F. Khan et al.

Intelligent segmentation of medical images using fuzzy bitplane thresholding

Meas Sci Rev

(2014)

T. Hachaj et al.

CAD system for automatic analysis of CT perfusion maps

Opto-Electron Rev

(2011)

T. Hachaj

Pattern classification methods for analysis and visualization of brain perfusion CT maps

Computational intelligence paradigms in advanced pattern classification

(2012)

Cited by (45)

A novel end-to-end deep convolutional neural network based skin lesion classification framework
2024, Expert Systems with Applications
Skin diseases are reported to contribute 1.79% of the global burden of disease. The accurate diagnosis of specific skin diseases is known to be a challenging task due, in part, to variations in skin tone, texture, body hair, etc. Classification of skin lesions using machine learning is a demanding task, due to the varying shapes, sizes, colors, and vague boundaries of some lesions. The use of deep learning for the classification of skin lesion images has been shown to help diagnose the disease at its early stages. Recent studies have demonstrated that these models perform well in skin detection tasks, with high accuracy and efficiency.
Our paper proposes an end-to-end framework for skin lesion classification, and our contributions are two-fold. Firstly, two fundamentally different algorithms are proposed for segmenting and extracting features from images during image preprocessing. Secondly, we present a deep convolutional neural network model, S-MobileNet that aims to classify 7 different types of skin lesions.
We used the HAM10000 dataset, which consists of 10000 dermatoscopic images from different populations and is publicly available through the International Skin Imaging Collaboration (ISIC) Archive. The image data was preprocessed to make it suitable for modeling. Exploratory data analysis (EDA) was performed to understand various attributes and their relationships within the dataset. A modified version of a Gaussian filtering algorithm and SFTA was applied for image segmentation and feature extraction. The processed dataset was then fed into the S-MobileNet model. This model was designed to be lightweight and was analyzed in three dimensions: using the Relu Activation function, the Mish activation function, and applying compression at intermediary layers. In addition, an alternative approach for compressing layers in the S-MobileNet architecture was applied to ensure a lightweight model that does not compromise on performance.
The model was trained using several experiments and assessed using various performance measures, including, loss, accuracy, precision, and the F1-score. Our results demonstrate an improvement in model performance when applying a preprocessing technique. The Mish activation function was shown to outperform Relu. Further, the classification accuracy of the compressed S-MobileNet was shown to outperform S-MobileNet.
To conclude, our findings have shown that our proposed deep learning-based S-MobileNet model is the optimal approach for classifying skin lesion images in the HAM10000 dataset. In the future, our approach could be adapted and applied to other datasets, and validated to develop a skin lesion framework that can be utilized in real-time.
A Novel Minutiae Triangulation Technique for Non-invertible Fingerprint Template Generation
2021, Expert Systems with Applications
Citation Excerpt :
So, the pre-processing and spurious minutiae elimination technique can also be explored. The feature extracted from histogram of image of finger skin proposed by Glowacz and Glowacz using image filtration can also be used with fingerprint feature to make the template revocable also (Glowacz & Glowacz, 2016). Amit Kumar Trivedi: Conceptualization, Methodology, Software, Writing - original draft, Reviewing, Investigation.
The fingerprint is the readily available and most widely used biometrics for the security of a system. In order to impart the security of the template database, it is important to make the fingerprint template robust to fingerprint reconstruction, i.e., invertibility. In this paper, a fingerprint template is proposed using a novel minutiae triangulation technique. In this triangulation technique, the set of minutiae of a template is taken to generate a set of minutiae triplets each forming a triangle. After this, a minutiae transformation is applied to obtain a set of triplets of internal angles of the triangle which also removes the minutiae location and orientation information. In the absence of this information, the orientation map of the fingerprint cannot be estimated. This set of triplets together with some other information is used as a proposed fingerprint template. The proposed system is experimented using standard FVC2000, FVC2004 and FVC2006 databases and yields better results in terms of EER (Equal Error Rate) as compared with the state-of-art techniques.
A new approach for computer-aided detection of coronavirus (COVID-19) from CT and X-ray images using machine learning methods
2021, Applied Soft Computing
The COVID-19 outbreak has been causing a global health crisis since December 2019. Due to this virus declared by the World Health Organization as a pandemic, the health authorities of the countries are constantly trying to reduce the spread rate of the virus by emphasizing the rules of masks, social distance, and hygiene. COVID-19 is highly contagious and spreads rapidly globally and early detection is of paramount importance. Any technological tool that can provide rapid detection of COVID-19 infection with high accuracy can be very useful to medical professionals. The disease findings on COVID-19 images, such as computed tomography (CT) and X-rays, are similar to other lung infections, making it difficult for medical professionals to distinguish COVID-19. Therefore, computer-aided diagnostic solutions are being developed to facilitate the identification of positive COVID-19 cases. The method currently used as a gold standard in detecting the virus is the Reverse Transcription Polymerase Chain Reaction (RT-PCR) test. Due to the high false-negative rate of this test and the delays in the test results, alternative solutions are sought. This study was conducted to investigate the contribution of machine learning and image processing to the rapid and accurate detection of COVID-19 from two of the most widely used different medical imaging modes, chest X-ray and CT images. The main purpose of this study is to support early diagnosis and treatment to end the coronavirus epidemic as soon as possible. One of the primary aims of the study is to provide support to medical professionals who are most worn out and working under intense stress during COVID-19 through smart learning methods and image classification models. The proposed approach was applied to three different public COVID-19 data sets and consists of five basic steps: data set acquisition, pre-processing, feature extraction, dimension reduction, and classification stages. Each stage has its sub-operations. The proposed model performs in considerable levels of COVID-19 detection for dataset-1 (CT), dataset-2 (X-ray) and dataset-3 (CT) with the accuracy of 89.41%, 99.02%, 98.11%, respectively. On the other hand, in the X-ray data set, an accuracy of 85.96% was obtained for COVID-19 (+), COVID-19 (-), and those with Pneumonia but not COVID-19 classes. As a result of the study, it has been shown that COVID-19 can be detected with a high success rate in about less than one minute with image processing and classical learning methods. In the light of the findings, it is possible to say that the proposed system will help radiologists in their decisions, will be useful in the early diagnosis of the virus, and can distinguish pneumonia caused by the COVID-19 virus from the pneumonia of other diseases.
Forecasting sidewall displacement of underground caverns using machine learning techniques
2021, Automation in Construction
Citation Excerpt :
For instance, a typical weighting pattern comprises giving each neighbor a weight of 1/d, where d is the space to the neighbor. KNN is a machine learning method type of instance-based learning or lazy learning, it only approximates function locally, and it defers all computation until evaluating the function [23,24]. The accuracy of normalizing the training data can be improved because this algorithm depends on the distance for classification [25].
Sidewall displacement is a vital task in underground caverns that can jeopardize the construction of these structures at risk and predict this event to provide the appropriate and timely solutions to deal with it can be very important. This article aims to predict sidewall displacement of the underground caverns using six machine learning approaches of long short-term memory, deep neural networks, K-nearest neighbors, Gaussian process regression, support vector regression, and decision tree. To this end, 310 datasets, including 12 useful parameters on the caverns' sidewall displacement, were gathered from 10 underground cavern projects in Iran and other countries. 74% of the datasets were applied to train the prediction models, and 26% for testing. Finally, through a comparison made between the prediction results and the measured ones, the most accurate predictions were produced by the LSTM model with correlation coefficient = 0.9998, root mean squared error = 0.002059 m, and variance account for = 99.98181%.
Detection of lung nodule and cancer using novel Mask-3 FCM and TWEDLNN algorithms
2021, Measurement: Journal of the International Measurement Confederation
Lung Cancer (LC) is reported as common cause of death all over the world. The detection of cancer can save many lives and help likelihood of survival. Physicians use CT (computed tomography) for examination of the cancer in lung with the help of computer-aided diagnosis (CAD) for efficient detection and diagnosis. The CAD uses different machine learning techniques and signal processing approaches but processing time and accuracy of CAD remains challenging issues. An efficient Deep Learning (DL) methodology is proposed for lung cancer detection utilizing Target based Weighted Elman DL Neural Network (TWEDLNN), and Mask Unit (MU) based 3FCM algorithm. The proposed work includes lung image segmentation using Geometric mean-based Otsu Thresholding (GOT); contrast enhancement (CE) using Modified Clip limit-based Contrasts Limited Adaptive Histograms Equalization (MC-CLAHE); Feature Extraction (FE); Classification of Features using TWEDLNN; and MU based FCM algorithm for LN (lung nodule) detection. We have used CT images of LIDC-IDRI database for the implementation. We have compared the proposed work with existing techniques to confirm that the TWEDLNN detects LC more efficiently and the accuracy of proposed work is also improved as 96%. The performance of proposed MC-CLAHE is authenticated by contrasting the proposed technique’s performance with prevailing techniques, CLAHE, Gaussian, Median, and Wiener filters. The proposed method has resulted PSNR of 24.2573 and MSE value of 292.98, which are better than all existing Techniques.
A novel fuzzy curvature method for recognition of anterior forearm subcutaneous veins by thermal imaging
2019, Expert Systems with Applications
Citation Excerpt :
The directional curvature methodology is only feasible for online recognition of the vein system by hand-held devices that will identify the directional veins while moving through a predefined direction. Moreover, there are a few recently published papers worth mentioning on biomedical or medical applications of thermal imaging which are Fu, Zeng, Ramírez-Pérez, & Lisak (2018), Majtner, Yildirim-Yayilgan, & Hardeberg (2016), Glowacz & Glowacz (2016), Koprowski (2016), Marzec, Koprowski, & Wróbel (2015), Snekhalatha, Rajalakshmi, Gopikrishnan, & Gupta (2017) and Ko et al. (2018). Marzec et al. (2015) proposed a face localization methodology determining head centers in thermograms without any further analysis; yet still relevant with our research for forearm segmentation.
Anterior forearm recognition systems emerged in last decades to identify the vein systems and to decide the venipuncture sites. Basically, identification and also real-time visualization of the forearm veins are commonly accomplished by near-infrared (NIR) camera systems in the literature and also in applied medicine; however what we propose in this paper is easier and reliable alternative by thermal imaging. While identifying vein systems, the common drawback of visible spectrum and NIR camera solutions is lack of recognition possibility of rather hidden veins in the forearms. In other words, while these solutions are so useful for identification of superficial veins like Median Cubital and Median Antebrachial veins; yet they are not so efficient for subcutaneous veins like Cephalic vein. Therefore, we introduce a novel fuzzy directional curvature methodology to recognize the whole vein system of anterior forearm using infrared thermal (IR-T) imaging. Initially, the forearm image captured by a thermal camera is segmented by crisp 2-means and filtered by Gaussian high-pass filter for smoothing and contrast enhancement. Four types of directional curvatures, achieved by second order derivatives including vertical, horizontal and two diagonal directions, are reproduced as four single images from scratch. The images are subsequently fused by fuzzy curvature method to produce the final images representing the complete vein system in the forearm. The fusion procedure by fuzzy inference system as an expert system could be stated as the main novelty in infrared thermal imaging which is also so flexible thanks to the parametric design. The intelligent recognition system also provides various clear screenings of the whole vein system in forearms depending on the parameters selected, even though the veins are totally invisible.

View all citing articles on Scopus

View full text

Original Research ArticleRecognition of images of finger skin with application of histogram, image filtration and K-NN classifier

Abstract

Introduction

Section snippets

Proposed method of recognition of image of finger skin

The results of recognition of images of injured human finger

Discussion and conclusions

Acknowledgments

Biocybern Biomed Eng

Neurocomputing

Comput Biol Med

Biocybern Biomed Eng

J Am Acad Dermatol

Comput Methods Progr Biomed

Biomed Signal Process Control

Performance evaluation in medical image segmentation

Curr Med Imaging Rev

A survey on medical image segmentation