ACCESSIBLE MELANOMA DETECTION USING SMARTPHONES AND MOBILE IMAGE ANALYSIS

Abstract

We investigate the design of an entire mobile imaging system for early detection of melanoma. Different from previous work, we focus on smartphone-captured visible light images. Our design addresses two major challenges. First, images acquired using a smartphone under loosely-controlled environmental conditions may be subject to various distortions, and this makes melanoma detection more difficult. Second, processing performed on a smartphone is subject to stringent computation and memory constraints. In our work, we propose a detection system that is optimized to run entirely on the resource constrained smartphone. Our system intends to localize the skin lesion by combining a lightweight method for skin detection with a hierarchical segmentation approach using two fast segmentation methods. Moreover, we study an extensive set of image features and propose new numerical features to characterize a skin lesion. Furthermore, we propose an improved feature selection algorithm to determine a small set of discriminative features used by the final lightweight system. In addition, we study the human-computer interface (HCI) design to understand the usability and acceptance issues of the proposed system. Our extensive evaluation on an image dataset provided by National Skin Center - Singapore (117 benign nevi and 67 malignant melanoma) confirms the effectiveness of the proposed system for melanoma detection: 89.09% sensitivity at specificity _ 90%.

Existing System

Landmark recognition, mobile video type classification and 3-D scene video method.

Proposed System

We propose a light-weight skin lesion localization algorithm suitable for the resource-constrained smartphone. Our localization algorithm comprises skin / non-skin detection, hierarchical segmentation, and combination of Otsu’s method and Minimum Spanning Tree (MST) method. We use novel color and border features to quantify the color variation and the border irregularity of skin lesions. We evaluate 116 computational features to quantify the color, border, asymmetry and texture of a skin lesion, including our proposed features that are suitable for visible light images captured under loosely-controlled lighting conditions. We investigate feature selection to identify a small set of the most discriminative features to be used in the smartphone. Using a small set of discriminative features not only reduces the storage and computation overhead but also improves the classification performance, as low dimensional feature vector is more robust to over-fitting. We focus on the framework using normalized mutual information and propose an improvement that takes into account the feature coordinates. We propose several methods to fuse the classification results of individual category classifiers. We evaluate our system using a dataset from National Skin Center (NSC) of Singapore. We study the Human Computer Interface (HCI) aspect of the proposed system.

Architecture

BLOCK DIAGRAM