Lesion classification using 3D skin surface tilt orientation Zhishun She 1,2 and P. S. Excell 1,2 1 Institute of Arts, Science & Technology, Glyndwr University, Wrexham, UK, and 2 University of Wales, Cardiff, UK Background/purpose: Current non-invasive diagnostic proce- dures 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 clini- cal (WLC) skin images by high-pass filtering. Then the direc- tions 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 ori- entation 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), dem- onstrated 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 orienta- tion (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 classi- fier accuracy. Combined with the ABCD features, it is very promising to distinguish malignant melanoma from benign lesions. Key words: 3D skin – tilt orientation – lesion classification – skin lines – melanoma Ó 2012 John Wiley & Sons A/S Accepted for publication 26 April 2012 M ALIGNANT MELANOMA is the most fatal type of skin cancer. As detection of malignant melanoma at an early stage considerably reduces its morbidity and patients’ mortality, computer automatic diagnosis (CAD) of skin lesions using early symptoms would be particu- larly useful as an aid in primary care. In order to implement this, a feature set enabling accu- rate differentiation between benign and malig- nant skin lesions is required. Most commonly used image-based diagnostic features are related to shape, boundary irregularity, colour variegation and size of skin lesion: the so-called ABCD features (1). These features are extracted from two-dimensional (2D) images of skin, but because the skin surface is an object in three- dimensional (3D) space, valuable additional information can be gained by investigating new features, which are derived from a consider- ation of 3D skin objects. Research on acquisition of 3D data and 3D image analysis of skin lesions and wounds has been reported. In 1984, Dhawan et al. (2) cap- tured three images to conduct 3D tomographic imaging. The vertical cross sections of the lesions were calculated to measure their thick- ness. In 1995, Jones and Plassmann (3) built an instrument based on the principle of colour- coded structured light. A set of parallel stripes was projected on a skin wound and a depth map was deduced from the distortions of the projected stripes. In 2009, Treuillet et al. (4) made use of two images to build 3D models of skin wounds using a structure from motion (SFM) algorithm. However, the computational cost of these methods is high for multiple image calibration and 3D image reconstruction. Recently, the photometric stereo method has been used to detect skin cancer. More than two images are acquired from different illumination 1 Skin Research and Technology 2012; 0:1–7 Printed in Singapore Á All rights reserved doi: 10.1111/j.1600-0846.2012.00644.x © 2012 John Wiley & Sons A / S Skin Research and Technology