Blood Stain Segmentation

Blood stain analysis is not automated and tedious for researchers. This paper proposes a method to segment blood stains from a laboratory generated blood spatter pattern to begin to automate this task. The images are converted into binary images by firstly improving the contrast in the image by converting it to a HSV image, then using Otsu thresholding to make the stains white and the background black. The stains are now detected using the Suzki border algorithm. An ellipse is then fitted to the detected areas for future processing. This resulted in all medium to large stains being identified with up to 17 percent false positives and up to 17 percent not identified stains. This is also robust against differences in lighting conditions and completes in under 6 seconds for 1814 stains identified. This is an improvement upon previous research which is much slower and only preforms well with tailored photographs.

[1]  S. N. Kunz,et al.  Blood stain pattern analysis , 2011, Forensic science, medicine, and pathology.

[2]  Pattara Aiyarak,et al.  Impact angle analysis of bloodstains using a simple image processing technique , 2010 .

[3]  Josef Kittler,et al.  A Comparative Study of Hough Transform Methods for Circle Finding , 1989, Alvey Vision Conference.

[4]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[5]  G. Zack,et al.  Automatic measurement of sister chromatid exchange frequency. , 1977, The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society.

[6]  Paul Suetens,et al.  Calculation of bloodstain impact angles using an Active Bloodstain Shape Model , 2014 .

[7]  J. A. Ojo,et al.  Contrast enhancement algorithm for colour images , 2015, 2015 Science and Information Conference (SAI).

[8]  Keiichi Abe,et al.  Topological structural analysis of digitized binary images by border following , 1985, Comput. Vis. Graph. Image Process..

[9]  Richard O. Duda,et al.  Use of the Hough transformation to detect lines and curves in pictures , 1972, CACM.

[10]  Henk J. A. M. Heijmans,et al.  The algebraic basis of mathematical morphology. I Dilations and erosions , 1990, Comput. Vis. Graph. Image Process..

[11]  Theresa Stotesbury,et al.  Investigation of Physical Effects of Acid Yellow 7® Enhancement on Dark and Non-Porous Surfaces in Impact Pattern Area of Origin Estimation , 2012 .

[12]  Michael C Taylor,et al.  An image-processing methodology for extracting bloodstain pattern features. , 2017, Forensic science international.

[13]  Dana H. Ballard,et al.  Generalizing the Hough transform to detect arbitrary shapes , 1981, Pattern Recognit..

[14]  Andrew W. Fitzgibbon,et al.  A Buyer's Guide to Conic Fitting , 1995, BMVC.

[15]  John F. Canny,et al.  A Computational Approach to Edge Detection , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[16]  Michael C Taylor,et al.  A novel, element-based approach for the objective classification of bloodstain patterns. , 2015, Forensic science international.