Thai Herb Leaves Classification Based on Properties of Image Regions

In this paper, we proposed a classification of Thai herb leaves based on the properties of image regions. The image regions that we considered are the leaves of herb images that placed over the smooth white background. The proposed method can be divided into three parts which are pre-processing, feature extraction, and classification. The pre-processing has performed to get the binary image from the input color of the herb leaves image. It composes of RGB to gray conversion, Histogram equalization, edge detection, and morphological techniques. After that, the binary image of Thai herb leaves is extracted with feature extraction using the region properties which are majority axis, minority axis, aspect ratio, eccentricity, circularity, extent, perimeter, and the ratio of the perimeter and majority axis. After obtained the feature vectors, an artificial neural network is executed to help classification. The experimental results demonstrate that the classification performance of the proposed method in term of accuracy achieved 90.50% on average.