Automatic dirt trail analysis in dermoscopy images

Beibei Cheng, R. Joe Stanley, William V. Stoecker, Christopher T.P. Osterwise, Sherea M. Stricklin, Kristen A. Hinton, Randy H. Moss, Margaret Oliviero, Harold S. Rabinovitz

Research output: Contribution to journalArticlepeer-review

9 Scopus citations


Background: Basal cell carcinoma (BCC) is the most common cancer in the US. Dermatoscopes are devices used by physicians to facilitate the early detection of these cancers based on the identification of skin lesion structures often specific to BCCs. One new lesion structure, referred to as dirt trails, has the appearance of dark gray, brown or black dots and clods of varying sizes distributed in elongated clusters with indistinct borders, often appearing as curvilinear trails. Methods: In this research, we explore a dirt trail detection and analysis algorithm for extracting, measuring, and characterizing dirt trails based on size, distribution, and color in dermoscopic skin lesion images. These dirt trails are then used to automatically discriminate BCC from benign skin lesions. Results: For an experimental data set of 35 BCC images with dirt trails and 79 benign lesion images, a neural network-based classifier achieved a 0.902 are under a receiver operating characteristic curve using a leave-one-out approach. Conclusion: Results obtained from this study show that automatic detection of dirt trails in dermoscopic images of BCC is feasible. This is important because of the large number of these skin cancers seen every year and the challenge of discovering these earlier with instrumentation.

Original languageEnglish (US)
Pages (from-to)e20-e26
JournalSkin Research and Technology
Issue number1
StatePublished - Feb 2013
Externally publishedYes


  • Basal cell carcinoma
  • Dermoscopy
  • Dirt trails
  • Image analysis
  • Neural network

ASJC Scopus subject areas

  • Dermatology


Dive into the research topics of 'Automatic dirt trail analysis in dermoscopy images'. Together they form a unique fingerprint.

Cite this