Cascade-forward neural network performance study for bloodstain image analysis

Cascade-Forward Neural Network (CFNN) performance is explored in this paper for blood stain image analysis. The blood stain images of various size, shape and impact angles are captured through experimentation. Each blood stain in the image is first detected using sobel edge detector. After the image has been thresholded and the noise removed, geometric properties of the blood drop is measured with `regionprops'. Regionprops is used to extract basic characteristics from acquired bloodstain images. These values are compiled into appropriate input to feed into the developed CFNN module for feature analysis and pattern recognition. Several trials have been conducted to determine the performance. In average the testing results show that CFNN is able to produce approximately 83.3% accuracy for blood stain image classifications. Hence, this method is simple yet effective for blood pattern analysis in forensic investigations.