Lesion classification using 3D skin surface tilt orientation.

BACKGROUND/PURPOSE: Current non-invasive diagnostic procedures to detect skin cancer rely on two-dimensional (2D) views of the skin surface. For example, the most commonly-used ABCD features are extracted from the 2D images of skin lesion. However, because the skin surface is an object in three-dimensional (3D) space, valuable additional information can be obtained from a perspective of 3D skin objects. The aim of this work is to discover the new diagnostic features by considering 3D views of skin artefacts. METHODS: A surface tilt orientation parameter was proposed to quantify the skin and the lesion in 3D space. The skin pattern was first extracted from simply captured white light optical clinical (WLC) skin images by high-pass filtering. Then the directions of the projected skin lines were determined by skin pattern analysis. Next the surface tilt orientations of skin and lesion were estimated using the shape from texture technique. Finally the difference of tilt orientation in the lesion and normal skin areas, combined with the ABCD features, was used as a lesion classifier. RESULTS: The proposed method was validated by processing a set of images of malignant melanoma and benign naevi. The scatter plot of classification using the feature of surface tilt orientation alone showed the potential of the new 3D feature, enclosing an area of 0.78 under the ROC curve. The scatter plot of classification, combining the new feature with the ABCD features by use of Principal Component Analysis (PCA), demonstrated an excellent separation of benign and malignant lesions. An ROC plot for this case enclosed an area of 0.85. Compared with the ABCD analysis where the area under the ROC curve was 0.65, it indicated that the surface tilt orientation (3D information) was able to enhance the classification results significantly. CONCLUSIONS: The initial classification results show that the surface tilt orientation has a potential to increase lesion classifier accuracy. Combined with the ABCD features, it is very promising to distinguish malignant melanoma from benign lesions.

Skin pattern analysis for lesion classification using local isotropy.

BACKGROUND/PURPOSE: The observation that skin pattern tends to be disrupted by malignant skin lesions, but not by benign ones suggests that measurements of skin pattern disruption on simply captured white light optical clinical (WLC) skin images could be a useful contribution to a diagnostic feature set. Previous work, which generated a flow field of skin pattern using a measurement of local line direction and intensity, was encouraging. The aim of this paper is to investigate the possibility of extracting new features using local isotropy metrics to quantify the skin pattern disruption. METHODS: The skin pattern was extracted from WLC skin images by high-pass filtering. A local tensor matrix was computed. The local isotropy was measured by the condition number of the local tensor matrix. The difference of this measure over the lesion and normal skin areas, combined with the local line direction and the ABCD features, was used as a lesion classifier. RESULTS: A set of images of malignant melanoma and benign naevi was analysed. A one-dimensional scatter plot showed the potential of a local isotropy metric, showing an area of 0.70 under the receiver operating characteristic (ROC) curve. A two-dimensional scatter plot, combined with the local line direction, indicated enhancement of the classification performance, showing an area of 0.89 under the ROC curve. A three-dimensional scatter plot combined with the local line direction and the ABCD features, using principal component analysis, demonstrated excellent separation of benign and malignant lesions. An ROC plot for this case enclosed an area of 0.96. CONCLUSION: The experimental results show that the local isotropy metric has a potential to increase lesion classifier accuracy. Combined with the local line direction and the ABCD features, it is very promising as a method to distinguish malignant melanoma from benign lesions.

Combination of features from skin pattern and ABCD analysis for lesion classification.

BACKGROUND/PURPOSE: It is known that the standard features for lesion classification are ABCD features, that is, asymmetry, border irregularity, colour variegation and diameter of lesion. However, the observation that skin patterning tends to be disrupted by malignant but not by benign skin lesions suggests that measurements of skin pattern disruption on simply captured white light optical skin images could be a useful contribution to a diagnostic feature set. Previous work using both skin line direction and intensity for lesion classification was encouraging. But these features have not been combined with the ABCD features. This paper explores the possibility of combing features from skin pattern and ABCD analysis to enhance classification performance. METHODS: The skin line direction and intensity were extracted from a local tensor matrix of skin pattern. Meanwhile, ABCD analysis was conducted to generate six features. They were asymmetry, border irregularity, colour (red, green and blue) variegations and diameter of lesion. The eight features of each case were combined using a principal component analysis (PCA) to produce two dominant features for lesion classification. RESULTS: A larger set of images containing malignant melanoma (MM) and benign naevi were processed as above and the scatter plot in a two-dimensional dominant feature space showed excellent separation of benign and malignant lesions. An ROC (receiver operating characteristic) plot enclosed an area of 0.94. CONCLUSIONS: The classification results showed that the individual features have a limited discrimination capability and the combined features were promising to distinguish MM from benign lesion.

Simulation and analysis of optical skin lesion images.

BACKGROUND/PURPOSE: In order to properly analyse the effectiveness of methods for optically differentiating malignant from benign skin lesions, it is necessary to have a set of images for which the ground truth is known. However, aspects of the ground truth of clinical images such as true lesion boundary position are unknown or not known precisely. Therefore, a skin/lesion image simulation with known features including boundary location, skin pattern and lesion colour is needed to enable accurate assessment of feature estimation algorithms for lesion classification. METHODS: In this paper, monochrome and colour skin/lesion images are synthesised with known characteristics including boundary, colour and skin pattern. Skin pattern is simulated with segmented lines with variations in length, orientation and intensity. Skin and lesion textures are modelled by an auto-regressive (AR) process. Monochrome skin lesion images are obtained by combining monochrome skin and lesion textures under the control of a known lesion shape with the addition of skin pattern. Colour skin lesion images are generated by mixing coloured skin and lesion textures. Finally, an inflammation area and image artefacts such as hair and specular reflection are added. RESULTS: The synthesised images provide the image set for evaluating image pre-processing, segmentation and skin pattern analysis. The pre-processing includes hair removal and specular reflection reduction. An AR model interpolation is suggested for hair removal, and multiple illumination processing is developed to decrease specular reflection. A fast snake algorithm is extended to detect the boundaries of skin lesion and inflammation areas. Skin line direction is detected as a feature to measure the disruption of skin pattern caused by lesion. CONCLUSIONS: Simulation of monochrome and colour skin/lesion image has been investigated, which is an alternative way to provide image set with known characteristics to validate image processing algorithms for image pre-processing, lesion/inflammation boundary detection and skin pattern analysis.

Analysis of skin line pattern for lesion classification.

BACKGROUND/PURPOSE: It has been observed that skin patterning tends to be disrupted by malignant but not by benign skin lesions. This suggests that measurements of skin pattern disruption on simply captured white light optical skin images could be a useful contribution to a diagnostic feature set. Previous work using a measurement of line strength by a consistent high-value profiling technique followed by local variance measurement or a region agglomerative classifier to measure skin line pattern disruption was extremely promising but computationally intensive, suggesting that the idea of measuring skin pattern disruption was useful but a simpler method was required. METHODS: The skin pattern was extracted by high-pass filtration and enhanced by adaptive anisotropic (spatial variant) filtering which smoothes along skin lines but not across them. The skin line main direction and direction variance were estimated using a local image gradient matrix and the difference of these measures across the lesion image boundary was used as a lesion classifier. RESULTS: A set of images of malignant melanoma and benign naevi were processed as above and the scatter plot of results in a two-dimensional feature (line direction and line variation difference) space showed excellent separation of benign and malignant lesions. An ROC plot enclosed an area of 0.88. CONCLUSIONS: The experimental results showed that the local line direction and the local line variation were promising features for distinguishing malignant melanoma from benign lesion and the methods used were effective and computationally low-cost.

Enhancement of lesion classification using divergence, curl and curvature of skin pattern.

BACKGROUND/PURPOSE: The observation that skin pattern tends to be disrupted by malignant but not by benign skin lesions suggests that measurements of skin pattern disruption on simply captured white light optical clinical (WLC) skin images could be a useful contribution to a diagnostic feature set. Previous work which generated a flow field of skin pattern using a measurement of local line direction and variation determined by the minimum eigenvalue and its corresponding eigenvector of the local tensor matrix to measure skin pattern disruption was computationally low cost and encouraging. This paper explores the possibility of extracting new features from the first and second differentiations of this flow field to enhance classification performance. METHODS: The skin pattern was extracted from WLC skin images by high-pass filtering. The skin line direction was estimated using a local image gradient matrix to produce a flow field of skin pattern. Divergence, curl, mean and Gaussian curvatures of this flow field were computed from the first and second differentiations of this flow field. The difference of these measures combined with skin line direction across the lesion image boundary was used as a lesion classifier. RESULTS: A set of images of malignant melanoma and benign naevi were analysed as above and the scatter plot in a two-dimensional dominant feature space using principal component analysis showed excellent separation of benign and malignant lesions. A receiver operating characteristic plot enclosed an area of 0.96. CONCLUSIONS: The experimental results show that the divergence, curl, mean and Gaussian curvatures of the flow field can increase lesion classifier accuracy. Combined with skin line direction they are promising features for distinguishing malignant melanoma from benign lesions and the methods used are computationally efficient which is important if their use is to be considered in clinical practice.