An Edge Extraction Method for Color Image Using Multiple-Valued LoG Filter and Color Space

All of the conventional edge extraction methods using LoG filter give output results having binary value. Therefore, the extracted edge always looses a lot of important information. To overcome this problem, the edge extraction method using LoG filter having multiple-valued output such as 3-valued and 5-valued has been proposed. In our work, to improve this extraction method so far, we emphasis the previous method by changing a size of the LoG filter and threshold values as the propose first method. Also, we propose the second method for getting a more detailed edge from a lightness information L, a chrominance information contained in I (orange-blue) information and Q (purple-green) information.