Two-dimensional signal and image processing
Two-dimensional signal and image processing
Texture Classification by Wavelet Packet Signatures
IEEE Transactions on Pattern Analysis and Machine Intelligence
Pattern Recognition, Fourth Edition
Pattern Recognition, Fourth Edition
Predictive network anomaly detection and visualization
IEEE Transactions on Information Forensics and Security
On the selection of an optimal wavelet basis for texture characterization
IEEE Transactions on Image Processing
| Hi-index | 0.00 |
This paper describes the various methods that are implemented to diagnose a sample of skin for malignancy. Skin cancer detection at the earliest stage possible is vital to increase the chance of survival of the affected patient. Imaging in this field happens to be at the cross-roads. Skin cancer imaging can be visual in nature (nevoscope imaging, electron microscope, naked eye) or non-visual (optical coherence tomography (OCT), Raman spectroscopy). ABCDs is a set of rules that are the first step that is applied to determine the nature of a mole. Although extensively used as front line methodology for malignancy in moles, it is not deterministic in nature. Each of the techniques described in this paper analyze the samples of the skin lesion under the scanner in a varied way. The samples of the skin lession can be either a visual depiction or in the form of a cross-section. We have after extensive experimentation arrived at two different ways to analyze the samples obtained as a result of the imaging. For the sample that we have obtained as a result of the nevoscope visual imaging, the power spectra appears to be the most discriminative and effective way of classification as against the use of discrete wavelet transformation in case of the cross-sections obtained from OCT. The aim is to ultimately build an automated system that has the capability to discriminating and classifying the skin samples into three main classes; namely, benign, precancerous and malignant independent of the scanning methodology.