Surface Imaging for Biomedical Applications PDF

Preface

As the editor and contributing author, I am motivated and compelled to write and compile this book after several clinical observational studies at Hospital Kuala Lumpur that investigated the use of signal and imaging processing techniques in dermatology for diagnostic and monitoring of skin diseases.

In clinical practice, dermatologists use both visual and tactile inspections to determine types and conditions of skin disorders. These inspection methods are highly subjective and thus require extensive training for use in clinical practice. Since 2004, we have been working with dermatologists in Hospital Kuala Lumpur on various skin disorders in developing objective measurement tools for diagnostic and monitoring purposes. Our measurement tools are developed based on image processing techniques combined with signal and statistical analyses, and are not only objective but also highly accurate.

In this book, we report the development of a psoriasis severity measurement tool called alpha-PASI that performs the Psoriasis Area Severity Index (PASI) gold standard that covers psoriasis lesion erythema, area, thickness, and scaliness. Several signal and image modalities are used. For example, 2D color data is used for determining lesion area, spectrophotometer data for lesion erythema while 3D surface imaging is used to determine thickness and scaliness of lesions.

The physician's global assessment of pigmentary skin disorders such as vitiligo, requires visual inspection by dermatologist, but pigmentation changes due to treatment take three to six months to discern visually. Therapeutic responses of vitiligo treatments are typically very slow and time consuming, and patients respond differently to treatments. Based on skin color model and using advanced image techniques with independent component analysis, the developed VT-Scan enables us to detect minute changes in pigmentation of the skin due to abnormal melanin production, thus reducing the interval between observations to several weeks. With VT-Scan, dermatologists are able to segment vitiligo areas and determine repigmentation areas accurately, allowing the assessment to be conducted within a shorter duration of six weeks compared to the typical three to six months.

The effectiveness of a treatment regime for chronic ulcers can be estimated by measuring changes in the ulcer wound. However, current ulcer management based on visual observation of the ulcer's conditions is not sufficient to determine treatment efficacy. Invasive methods for wound measurements are time consuming and often result in inconsistency of patient care. We have developed a volume ulcer assessment that uses non-invasive 3D imaging techniques to determine ulcer volume objectively. 3D laser scanning (laser triangulation) and optical scanning (structured light) techniques are investigated. Algorithms for solid modelling and volume computation from 3D surface scans are developed.

Most of the diagnosis process carried out by dermatologists for assessment of the acne vulgaris lesion is visual in nature. In grading acne vulgaris, the identification of acne lesion type and the counts for each particular type at various parts of the face, upper back and chest, is a time-consuming and tedious process. As a result, the dermatologists make an estimation of the type and the count of acne lesions and assign the grade based on that approximation; this introduces subjectivity and it also leads to inter- and intra-variability. We have developed an automated system for acne grading that is based on capturing the images of various body parts using the DSLR camera. Contrast enhancement and segmentation are two of the key and essential steps of the automated acne grading system. Contrast enhancement is achieved by artificially generating the high dynamic range images of acne lesions using the local rank transform followed by log-based tone mapping while the acne lesions segmentation method is based on the spectral models of acne lesions.

Lastly, to my fellow biomedical imaging researchers, I hope this book will bring out the passion of research that transcends disciplines and professions.