Cell detection in very low contrast images using Discrete Curvelet Transform and radon transform with morphological operations

Cell detection has been a crucial area in modern cell image processing applications. The low contrast cell images is a major limitation in cell detection. This paper proposes a method to detect cells in very low contrast cell images using Fast Discrete Curvelet Transform (FDCT), radon transform and morphological operations by reconstruction. The contrast of the cell images is improved by nonlinearly modifying the curvelet coefficients at selective scales. Further, radon transform is applied to reconstruct the image from the preprocessed image. Finally, the optimum morphological operations have been applied on the processed images to extract the cell regions from the low contrast cell images. The proposed method has been tested and improved cell detection results have been obtained.