Hough transformation based analysis technique for zebrafish embryo images

An analysis technique based on patched zebrafish embryo microscopic images was proposed to develop a computerized tool for zebrafish image analysis and quantitation. Three morphological features: lengths, head-trunk angles and tail curvatures of zebrafish embryos were obtained. First, morphological operations were used to segment the zebrafish embryo from the background. Then, its central line was extracted by using the binary image thinning algorithm. At last, three features were calculated. As for the tail curvatures, the modified circular Hough transformation was applied to remove the head of the fish, then the curved skeleton of the tail was regressed onto a straight line by using the least area method. The root mean squared error (RMSE) which represented the averaged distance between the skeleton curve and the regressed line was used as the third feature. Experimental results show that the computerized analysis technique has high accuracy and runs fast.