jdy (57K)

Jennifer G. Dy

Department of Electrical and Computer Engineering
Northeastern University

3-D Segmentation and Classification of Skin Images from Confocal Laser Scanning Microscope

  • Objective: Segmentation of dermis/epidermis boundary (basal layer) from confocal microscopy images

  • Why is it so important?  Because cancerous lesions begin at this layer first and then spread

  • Imaging modality: Confocal Scanning laser reflectance microscopy (CSLM):

    • Powerful in-vivo modality for 3-D imaging of skin (Z-stack of 2-D images from different z locations)

  • Different methods are applied for different skin types:

    • Fair skin has little melanin, hence brighter basal cells are not observed

    • Dark skin: noisy but the basal cells are more observable

  • Several techniques are applied for different skin types including:

    • Active Contours

    • Feature extraction and classification

This research is in collaboration with Prof. Dana Brooks (of Northeastern University, whose field is biomedical imaging), Dr. Milind Rajadhyaksha, Ph.D., (an expert in biomedical optics) and Dr. Allan Halpern, M.D. (a dermatologist) from the Dermatology Department of the Memorial Sloan-Kettering Cancer Center.

The student working on this project is: