Skin cancer detection methods
Pro Research Analysisby 
Searched over 200M research papers
Analyzed relevant papers
Overview of Skin Cancer Detection Methods
Skin cancer detection methods have evolved significantly, ranging from traditional visual inspection to advanced non-invasive imaging and artificial intelligence-based systems. Early and accurate detection is crucial for improving patient outcomes, especially for aggressive types like melanoma 347.
Traditional and Non-Invasive Diagnostic Techniques
Historically, skin cancer diagnosis relied on visual inspection and biopsy, which can be invasive, time-consuming, and sometimes lead to unnecessary procedures 369. To address these limitations, several non-invasive diagnostic tools have been developed, including:
- Photography and Dermoscopy: High-resolution imaging and dermoscopic analysis help visualize skin lesions in detail, aiding in early detection .
- Sonography and Confocal Microscopy: These imaging techniques provide deeper insights into skin structure without the need for tissue removal .
- Spectroscopy (Raman, Fluorescence, Terahertz): These methods analyze the chemical composition of skin lesions, improving diagnostic accuracy .
- Optical Coherence Tomography and Multispectral Imaging: These techniques offer cross-sectional and multi-wavelength views of skin, enhancing lesion characterization .
- Thermography and Dynamic Thermal Imaging (DTI): DTI, in particular, measures thermal recovery of lesions after cooling, achieving high sensitivity and specificity (>99%) for skin cancer detection 36.
- Electrical Bio-impedance and Tape Stripping: These methods assess electrical properties and collect superficial skin samples for analysis .
Computer-Aided and Artificial Intelligence-Based Detection
Automated systems using artificial intelligence (AI) and machine learning have become increasingly important for skin cancer detection, offering high accuracy and efficiency 1245+4 MORE.
Image Preprocessing and Feature Extraction
- Preprocessing: Techniques such as resizing, normalization, grayscale conversion, and edge detection (e.g., Canny edge detection) are used to prepare images for analysis 14.
- Feature Extraction: Key features include asymmetry, border irregularity, color, diameter, and texture. Methods like the ABCD rule, Histogram of Oriented Gradients (HOG), and Gray Level Co-occurrence Matrix (GLCM) are commonly used 14.
Machine Learning and Deep Learning Approaches
- Classical Machine Learning: Algorithms like Support Vector Machine (SVM), k-Nearest Neighbors (KNN), Naïve Bayes, and Random Forest are used to classify lesions as benign or malignant based on extracted features 159.
- Deep Learning: Convolutional Neural Networks (CNNs) and advanced architectures like ResNet50 and Inception V3 automatically learn complex patterns from images, improving detection accuracy 2457+2 MORE.
- Ensemble Methods: Combining multiple machine learning and deep learning models (ensemble classifiers) further enhances prediction accuracy and robustness 25.
Performance and Accuracy
- Automated systems have achieved high accuracy rates, with some models reporting up to 97.8% accuracy using SVM and 94% accuracy with Artificial Neural Networks (ANN) 19.
- Deep learning models, especially when optimized or combined with feature selection algorithms, consistently outperform traditional methods and even expert dermatologists in some cases 2510.
- Sensitivity and specificity values are also high, indicating reliable detection of both benign and malignant lesions 1259.
Advantages and Limitations
- Advantages: Non-invasive methods and AI-based systems reduce the need for biopsies, minimize patient discomfort, and enable early, accurate, and efficient diagnosis 369.
- Limitations: Some methods may require high-quality imaging, large datasets for training, and careful feature selection to avoid misdiagnosis or overfitting 1247+1 MORE.
Conclusion
Skin cancer detection methods have advanced from traditional visual inspection to sophisticated non-invasive imaging and AI-driven systems. Non-invasive techniques like dermoscopy, spectroscopy, and dynamic thermal imaging provide detailed lesion analysis, while machine learning and deep learning models offer high accuracy and efficiency in automated diagnosis. These advancements are crucial for early detection, reducing unnecessary biopsies, and improving patient outcomes 1234+6 MORE.
Sources and full results
Most relevant research papers on this topic
1
Skin Cancer Detection using Machine Learning Techniques
The proposed algorithm accurately detects skin lesions using ABCD rule, GLCM, and HOG features, achieving 97.8% accuracy and 85.1% sensitivity using SVM classifiers and 86.2 % sensitivity and 85.1% specificity using KNN classifiers.
Highly Cited
2020·
118citations
·V. M. et al.2020 IEEE International Conference on Electronics, Computing and Communication Technologies (CONECCT)·
DOI
2
Skin cancer detection using ensemble of machine learning and deep learning techniques
Our novel approach combining machine learning and deep learning techniques achieves a higher accuracy of 93% for skin cancer detection, outperforming expert dermatologists and other state-of-the-art methods.
Highly Cited
2023·
151citations
·Jitendra V. Tembhurne et al.3
Skin cancer detection using non-invasive techniques
Non-invasive skin cancer diagnostic methods, such as photography, dermoscopy, sonography, and confocal microscopy, offer promising alternatives to traditional invasive techniques, offering potential for improved detection and diagnosis.
Literature ReviewHighly Cited
2018·
119citations
·Vigneswaran Narayanamurthy et al.PDF
5
Skin cancer diagnosis: Leveraging deep hidden features and ensemble classifiers for early detection and classification
The proposed deep hidden feature and ensemble classifier skin cancer detection and classification technique improves prediction accuracy compared to state-of-the-art classifiers, enhancing early skin cancer diagnosis.
Highly Cited
2023·
80citations
·G. Akilandasowmya et al.6
Detection theory for accurate and non-invasive skin cancer diagnosis using dynamic thermal imaging.
Dynamic thermal imaging (DTI) provides a rapid, accurate, and non-invasive method for detecting skin cancer, with a sensitivity in excess of 99% and prescribed specificity in excess of 99%.
2017·
52citations
·S. E. Godoy et al.PDF
9
Early automated detection system for skin cancer diagnosis using artificial intelligent techniques
The proposed automated early detection system for skin cancer using artificial intelligence techniques shows high accuracy, precision, specificity, and sensitivity, enabling early detection and reduced mortality rates.
2024·
55citations
·Nourelhoda M. Mahmoud et al.PDF