The quality of digital image of eruption varies by photography condition, and influences the reliability of a result in fractal or other image processing. Hessian is a good mathematical element for these processing, and consists of quadratic differential elements provided from a three-dimensional figure of brightness on digital images. The picture processed by Hessian was evaluated through the several filters that extracted straight lines, squares, dots and curves in a limited part. In eczema there were many dot elements with a high simple fractal dimension. Large square compositions in lichen planus and whirlpools in Bowen disease were detected. A coefficient of variation of fractal dimension on Hessian picture was large in basal cell carcinoma and squamous cell carcinoma compared to actinic keratosis and seborrheic keratosis. Making of a real surface figure from a digital picture of eruption by Hessian-applied method was also tried. The surface change of the artificial epithelial growth in the computer memory was tested, and its dimensional change of a surface was also examined. Conclusion: A coefficient of variation of fractal dimension on divided Hessian picture is a good marker of abnormality of eruption compared with a simple fractal dimension.