抄録
This paper reports studies conducted on image processing with morphological methods. Morphological methods are based on mathematical morphology. These methods can analyze the spatial structure and shape of objects in a target image. We focused on the effectiveness of morphological methods for image processing on embedded processors. Several studies on this topic have been conducted and reported in [1]-[6]. The max-plus algebra-based morphological wavelet transform (MMT) is suitable for parallel processing on simple low-power embedded processors [1]. The adaptive multidirectional max-plus algebra-based morphological wavelet transform (AM-MMT) captures directional features of objects in an image for data compression [6]. MMT watermarking is a data-embedding algorithm for highly parallel processing [2]. The morphological pattern spectrum used for detecting image manipulation extracts information on target objects in an image and detects several image manipulations [3]-[5]. Research on these topics opens up potential for various applications of morphological methods. We plan to continue research on morphological-method-based image-processing algorithms for embedded processors.
