Conference Paper

Feature Selection Classification of Skin Cancer using Genetic Algorithm

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... A comprehensive analysis on 15 datasets showed that the proposed algorithm could reduce up to 70-80% of features and simultaneously provide better classification accuracy. On the other hand, Srividya and Arulmozhi (2018) explored the effects of FS powered by genetic algorithms on a case study of skin cancer patients, where a k-nearest neighbor classifier was exploited as a data mining method. A similar methodology, i.e., hybrid FS combining mutual information maximization and adaptive genetic algorithms, was also used for medical purposes (gene expression). ...
Article
Full-text available
Large information datasets often impose an immense number of features where many are found redundant and thus inessential for statistical analysis. In the past, a data preprocessing phase was formalized to cope with the problem and take appropriate remedial measures. Traditionally, this was a fixed and stationary process that suffered from a lack of transparency and high susceptibility to input variations. This paper presents a novel and fully automated meta-heuristic nature-inspired wrapper-based feature selection framework DynFS with dynamically cutting search space. The experiments show that the DynFS statistically significantly overcomes a fixed feature selection framework and allows for a high level of robustness and stability.
... Tahapan dull razor filtering untuk menghilangkan piksel rambut pada citra RGB sebagai berikut (Srividya and Arulmozhi, 2018 Bentuk umum dari matriks kookurensi adalah sebagai berikut: Bentuk umum dari matriks simetri adalah sebagai berikut: ...
Thesis
Full-text available
Kanker kulit merupakan salah satu dari jenis kanker yang menyerang lapisan kulit, penyakit ini disebabkan karena adanya radiasi dari sinar UV, faktor genetik dan toksin tertentu. Sel kanker kulit sendiri jika tidak mendapatka penanganan yang benar dapat berakibat fatal seperti kematian. Oleh karena itu, penelitian ini memeliki tujuan yaitu untuk mengklasifikasikan penyakit kanker kulit menggunakan Gray Level Co-occurrencence Matrix (GLCM) dan Kernel Extreme Learning Machine (KELM) dalam tiga kelas yaitu melanoma, actinic keratosis, dan nevus. Penelitian ini menggunakan data Melanoma sebanyak 144, Actinic Keratosis sebanyak 130, dan Nevus sebanyak 134. Tahapan yang dilakukan pada penelitian ini yaitu input data, kemudaian pre-processing dengan dull razor filtering, lalu data diubah ke bentuk grayscale. Setelah itu ekstraksi fitur dengan 5 fitur yaitu kontras, korelasi, energi, homogenitas, dan entropi. Hasil ekstraksi fitur yang diperoleh memberikan informasi bahwa, seluruh fitur dari kelas nevus beririsan dengan fitur kelas melanoma. Sedangkan pada fitur energi dan entropi nilai ketiga kelas tersebut saling beririsan. Pada tahapan klasifikasi, pembagian data training dan testing menggunakan K-fold cross validation dimana nilai k = 10. Metode yang digunakan pada tahapan klasifikasi yaitu metode menggunakan KELM. Uji coba yang dilakukan ada 3 uji coba, yaitu uji coba sudut GLCM, fungsi kernel, dan koefisien regulasi (c). Kemudian evaluasai model dengan confusion matrix. Hasil uji coba terbaik diperoleh dengan menggunakan GLCM pada sudut 45°, kernel yang digunakan yaitu kernel RBF dan wavelet, serta nilai c = 1. Nilai akurasi, sensitivitas, dan spesifisitas yaitu 95.12%, 95.24%, dan 97.53%.
Chapter
In healthcare sector, cancer is one of the most threatening and fast-growing diseases. The early diagnosis of this disease is very important as the success rate of its treatment depends upon how early and accurately it is diagnosed. The machine learning algorithms are helpful in detection and prediction of diseases. To improve efficiency of these algorithms, optimal features need to be selected. So, this research work uses genetic algorithm to select optimal features before applying k-nearest neighbor (KNN) and weighted k-nearest neighbor (WKNN) on Wisconsin Breast Cancer Prognosis dataset extracted from UCI repository. This approach helps in early prediction and the results show that WKNN performed better with 86.44% accuracy than KNN which gives 83.05% accuracy.
Article
Full-text available
Skin cancer is one of the most growing types and dangerous cancer in the world; the important of these cancers are malignant melanoma. The early diagnosis of malignant melanoma is a critical issue for dermatologists. In this paper, we present an overview of recent the state of the art in Computer-aided detection/diagnosis (CAD) systems in identifying and diagnosing malignant melanoma of dermoscopy images and describe its steps starting with image acquisition, preprocessing; and finishing with malignant melanoma classification of dermoscopic images. The comparative study shows that the most common methods for features extraction are the Discreet Wavelet Transform (DWT) and the method which combines both texture and color features resulting in output of very high accuracy. The methods for the classification:í µí°¾-Nearest Neighbor, Artificial Neural Networks, and Support Vector Machines are very well in the range [%90-% 97, 5].
Article
Full-text available
Dense object detection and temporal tracking are needed across applications domains ranging from people-tracking to analysis of satellite imagery over time. The detection and tracking of malignant skin cancers and benign moles poses a particularly challenging problem due to the general uniformity of large skin patches, the fact that skin lesions vary little in their appearance, and the relatively small amount of data available. Here we introduce a novel data synthesis technique that merges images of individual skin lesions with full-body images and heavily augments them to generate significant amounts of data. We build a convolutional neural network (CNN) based system, trained on this synthetic data, and demonstrate superior performance to traditional detection and tracking techniques. Additionally, we compare our system to humans trained with simple criteria. Our system is intended for potential clinical use to augment the capabilities of healthcare providers. While domain-specific, we believe the methods invoked in this work will be useful in applying CNNs across domains that suffer from limited data availability.
Article
Full-text available
The incidence of malignant melanoma continues to increase worldwide. This cancer can strike at any age; it is one of the leading causes of loss of life in young persons. Since this cancer is visible on the skin, it is potentially detectable at a very early stage when it is curable. New developments have converged to make fully automatic early melanoma detection a real possibility. First, the advent of dermoscopy has enabled a dramatic boost in clinical diagnostic ability to the point that melanoma can be detected in the clinic at the very earliest stages. The global adoption of this technology has allowed accumulation of large collections of dermoscopy images of melanomas and benign lesions validated by histopathology. The development of advanced technologies in the areas of image processing and machine learning have given us the ability to allow distinction of malignant melanoma from the many benign mimics that require no biopsy. These new technologies should allow not only earlier detection of melanoma, but also reduction of the large number of needless and costly biopsy procedures. Although some of the new systems reported for these technologies have shown promise in preliminary trials, widespread implementation must await further technical progress in accuracy and reproducibility. In this paper, we provide an overview of computerized detection of melanoma in dermoscopy images. First, we discuss the various aspects of lesion segmentation. Then, we provide a brief overview of clinical feature segmentation. Finally, we discuss the classification stage where machine learning algorithms are applied to the attributes generated from the segmented features to predict the existence of melanoma.
Conference Paper
Full-text available
In this paper, a novel approach for automatic segmentation and classification of skin lesions is proposed. Initially, skin images are filtered to remove unwanted hairs and noise and then the segmentation process is carried out to extract lesion areas. For segmentation, a region growing method is applied by automatic initialization of seed points. The segmentation performance is measured with different well known measures and the results are appreciable. Subsequently, the extracted lesion areas are represented by color and texture features. SVM and k-NN classifiers are used along with their fusion for the classification using the extracted features. The performance of the system is tested on our own dataset of 726 samples from 141 images consisting of 5 different classes of diseases. The results are very promising with 46.71% and 34% of F-measure using SVM and k-NN classifier respectively and with 61% of F-measure for fusion of SVM and k-NN.
Article
Full-text available
This paper presents an algorithm for classification of non-melanoma skin lesions based on a novel hierarchical K-Nearest Neighbors (K-NN) classifier. The K-NN classifier is simple, quick and effective. The hierarchical structure decomposes the classification task into a set of simpler problems, one at each node of the classification. Feature selection is embedded in the hierarchical framework that chooses the most relevant feature subsets at each node of the hierarchy. Colour and texture features are extracted from skin lesions. The accuracy of the proposed hierarchical scheme is higher than 93% in discriminating cancer and pre-malignant lesions from benign lesions, and it reaches an overall classification accuracy of 74% over five common classes of skin lesions, including two non-melanoma cancer types. This is the most extensive published result on non-melanoma skin cancer classification from colour images acquired by a standard camera (non-dermoscopy).
Article
Full-text available
This paper describes application of Genetic algorithm (GA) for design of network configuration and for learning of neural network. Design of network configuration is the first area for GA exercise in relation to neural network. The number of neurons in network and placement to the layers has big influence over effectivity of whole system. If we are able to formulate quality classification of designed network from standpoint of topology, we can use GA for design of suitable network configuration. The second area (learning of neural network) consists in using of advantages of GA toward learning of neural networks. In this case GA looks for acceptable setting of network weights so, to make specified transformation – it practices minimalization of its mistake function. The Genetic algorithm is considered to be a stochastic heuristic (or meta-heuristic) method. Genetic algorithms are inspired by adaptive and evolutionary mechanisms of live organisms. The best use of Genetic algorithm can be found in solving multidimensional optimisation problems, for which analytical solutions are unknown (or extremely complex) and efficient numerical methods are also not known.
Article
Full-text available
This paper presents a classifier that is based on a modified version of the well known K-Nearest Neighbors classifier (K-NN). The original K-NN classifier was adjusted to work with category representatives rather than training documents. Each category was represented by one document that was constructed by consulting all of its training documents and then applying feature selection so that only important terms remain. By this, when classifying a new document, it is required to be compared with category representatives and these are usually substantially fewer than training documents. This modified K-NN was experimented with in a hierarchical setting, i.e. when categories are represented as a hierarchy. Also, a new document similarity measure was proposed. It focuses on co-occurring or matching terms between a document and a category when calculating the similarity. This measure produces classification accuracy compared to the one obtained if the cosine, Jaccard or Dice similarity measures were used; yet it requires a much less time. The TrechTC-100 hierarchical dataset was used to evaluate the proposed classifier.
Conference Paper
Full-text available
Addresses design and evaluation of neural classifiers for the problem of skin lesion classification. By using Gauss Newton optimization for the entropic cost function in conjunction with pruning by Optimal Brain Damage and a new test error estimate, the authors show that this scheme is capable of optimizing the architecture of neural classifiers. Furthermore, error-reject tradeoff theory indicates, that the resulting neural classifiers for the skin lesion classification problem are near-optimal
Article
Full-text available
A large number of algorithms have been proposed for feature subset selection. Our experimental results show that the sequential forward floating selection algorithm, proposed by Pudil et al. (1994), dominates the other algorithms tested. We study the problem of choosing an optimal feature set for land use classification based on SAR satellite images using four different texture models. Pooling features derived from different texture models, followed by a feature selection results in a substantial improvement in the classification accuracy. We also illustrate the dangers of using feature selection in small sample size situations
Article
Full-text available
Needs of feature selection in medium and large problems (medium- and large-scale feature selection) increases in many fields including medical and image processing fields. Previous comparative studies of feature selection algorithms are not sufficient in problem size and in criterion function. In addition, no way has not shown to compare algorithms with different objectives. In this study, we propose a unified way to compare a large variety of algorithms. Our results on some medium and large problems show that the sequential floating algorithms promises for up to medium problems and genetic algorithms for medium and large problems.
Article
Computers have been used for mechanization and automation in different applications of agriculture/horticulture. The critical decision on the agricultural yield and plant protection is done with the development of expert system (decision support system) using computer vision techniques. One of the areas considered in the present work is the processing of images of plant diseases affecting agriculture/horticulture crops. The first symptoms of plant disease have to be correctly detected, identified, and quantified in the initial stages. The color and texture features have been used in order to work with the sample images of plant diseases. Algorithms for extraction of color and texture features have been developed, which are in turn used to train support vector machine (SVM) and artificial neural network (ANN) classifiers. The study has presented a reduced feature set based approach for recognition and classification of images of plant diseases. The results reveal that SVM classifier is more suitable for identification and classification of plant diseases affecting agriculture/horticulture crops.
Article
This paper proposes a method of genetic algorithm (GA) based neural network for feature selection that retains sufficient information for classification purposes. This method combines a genetic algorithm with an artificial neural network classifier, such as back-propagation (BP) neural classifier, radial basis function (RBF) classifier or learning vector quantization (LVQ) classifier. In this article, the genetic algorithm optimizes a feature vector by removing both irrelevant and redundant features and finds optimal ones. First, the procedure of the proposed algorithm is described and then the performance of this method is evaluated using two data sets. The results are compared with the genetic algorithm in combination with the k-nearest neighbor (KNN) classification rule. Our results suggest that GA based neural classifiers are robust and effective in finding optimal subsets of features from large data sets.
Article
The performance of any classification method relies on the quality of the measurement provided. One way to improve the classification is by extending the feature set and selecting the best features out of this set. In this paper a set of non-information-preserving features based on a multiresolution curve analysis which makes shape features explicit at multiple scales is proposed. An automatic procedure is designed to construct a binary tree which is used to classify the unknown objects.
NeuralNetwork Based Detection of Melanoma Skin Cancer
  • S Saudamini
  • Amit Jivtode
  • Ukalkar
SVM and ANN Based Classification Of Plant Diseases Using Feature Reduction Technique
  • D Jagadeesh
  • Rajesh Pujari
  • Yakkundimath
The Melanoma Skin Cancer Detection and Feature Extraction
  • S Gopinathan
  • S Nancy Arokia Rani
New Technique Of Genetic Algorithm For Finding Maximum Clique Problem
  • Savita Sudha
Review of Segmentation Techniques for Melanoma Detection
  • Ruchika Sharma
  • Pankaj Mohindru
Segmentation and Classification of Skin Lesions for Disease Diagnosis
  • Mahamad Sumithra Ra
  • D S Suhilb
  • Guruc