Papers
Topics
Authors
Recent
Assistant
AI Research Assistant
Well-researched responses based on relevant abstracts and paper content.
Custom Instructions Pro
Preferences or requirements that you'd like Emergent Mind to consider when generating responses.
Gemini 2.5 Flash
Gemini 2.5 Flash 134 tok/s
Gemini 2.5 Pro 41 tok/s Pro
GPT-5 Medium 30 tok/s Pro
GPT-5 High 26 tok/s Pro
GPT-4o 64 tok/s Pro
Kimi K2 185 tok/s Pro
GPT OSS 120B 442 tok/s Pro
Claude Sonnet 4.5 37 tok/s Pro
2000 character limit reached

Segmentation and Classification of Skin Lesions for Disease Diagnosis (1609.03277v1)

Published 12 Sep 2016 in cs.CV

Abstract: 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.

Citations (201)

Summary

Segmentation and Classification of Skin Lesions for Disease Diagnosis: An In-Depth Analysis

The paper "Segmentation and Classification of Skin Lesions for Disease Diagnosis" by R. Sumithra et al., presents a structured approach for the segmentation and classification of skin lesions using digital image processing techniques. The research highlights its applicability in the development of Computer-Aided Diagnosis Systems (CADs), which play a crucial role in assisting dermatologists, particularly those less experienced, by leveraging machine learning for automated diagnostics.

Methodology Overview

The paper introduces a sophisticated methodology encompassing several key phases: preprocessing, segmentation, feature extraction, and classification. The preprocessing stage involves noise reduction using Gaussian smoothing and morphological operations to enhance image quality by removing hairs and noise. This sets the groundwork for the subsequent segmentation phase, which employs a region growing method with automatic initialization of seed points. The performance of this segmentation approach is rigorously evaluated using measures such as Measure of Overlap (MOL), Measure of Under Segmentation (MUS), Measure of Over Segmentation (MOS), Dice Similarity Measure (DSM), and Error Rate (ER).

Following segmentation, the paper focuses on feature extraction. This process relies extensively on color and texture descriptors, including color moments across diverse color spaces (RGB, HSV, YCbCr, NTSc, CIE L*u*v, and CIE L*a*b) and texture features derived from the Gray Level Co-occurrence Matrix (GLCM). Together, these features form a comprehensive 4182-dimensional feature vector for each lesion.

Classification Techniques and Results

The classification phase utilizes Support Vector Machine (SVM) and k-Nearest Neighbors (k-NN) classifiers, with an additional layer of decision fusion using an OR rule to enhance classification performance. The efficacy of these techniques is evaluated using a custom dataset of 141 images resulting in 726 lesion samples across five skin disease classes: Melanoma, Bullae, Seborrheic Keratosis, Shingles, and Squamous Cell.

The results affirm the superiority of combined classifiers, achieving an F-measure of 61% with SVM and k-NN fusion compared to 46.71% and 34%, respectively, for the individual classifiers. The paper highlights the challenges posed by certain classes due to intraclass variance and interclass similarity, which affect the overall classification accuracy.

Implications and Future Directions

The research underscores the potential of automated diagnostic systems in dermatology, which could facilitate early and accurate detection of skin diseases. However, the current paper's limitations—primarily the dataset's uncontrolled acquisition parameters—suggest a need for further refinement. Future efforts could focus on acquiring standardized datasets and exploring feature selection techniques to enhance classification accuracy. Additionally, evaluating classifiers such as Artificial Neural Networks (ANNs) or Probabilistic Neural Networks (PNNs), and experimenting with multi-level fusion strategies (e.g., feature-level), may yield better performance outcomes.

The comparative analysis with existing methodologies in the paper reflects its competitive performance, establishing a foundation for additional research aimed at improving diagnostic reliability and scope in CAD systems focused on dermatological conditions.

Dice Question Streamline Icon: https://streamlinehq.com

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Lightbulb Streamline Icon: https://streamlinehq.com

Continue Learning

We haven't generated follow-up questions for this paper yet.

List To Do Tasks Checklist Streamline Icon: https://streamlinehq.com

Collections

Sign up for free to add this paper to one or more collections.