Edge detection in electron microscopy biological images using statistical dispersion

During the last few decades, there has been a tremendous development in the field of biological sciences. With this, there is an increasing demand for analyzing the molecular or cellular features of the cell in the images, acquired with the electron microscopes (EM). However, despite significant progress in image processing, the efficient detection of features and edges in biological images is still a challenging task due to the presence of minute structures with low intensity variation, compared with the background. In this paper, a novel algorithm for edge detection in electron microscopy biological image is proposed. The edge detector is based on the statistical dispersion of D8 pixels followed by an edge thinning operation. The proposed algorithm has been compared with other state-of-art edge detectors viz. Sobel's and Canny's edge detectors and results suggest that the proposed scheme perform better in detecting the significant edges in Tobacco Mosaic Virus (TMV; scanning-transmission electron microscopy image) and Virus Like Particles (VLPs; transmission electron microscopy image); used as test images in the present study. Experimental results (in terms of Pratt's figure of merit) also suggest that the proposed algorithm is more robust to noise when compared to Sobel's or Canny's edge detector. Consequently, the proposed algorithm can operate efficiently in a noisier environment, compared to Sobel's or Canny's edge detector.