ABSTRACT
54With the increasing demand of medicinal herbal plant globally every year, it is vital to preserve it using information technology so that the herbal medicinal plant community as well as next generation can certainly access the information efficiently. Instead of the common textual information, the medicinal herbal plant image is one of the main types of information that people always looking for. It is because human beings acquire the majority of information from the real world from their visual sense. However, to retrieve the required image for an ordinary user is a challenging task. User has a tendency to use high-level feature to interpret the images and measure the similarity rather than using low-level feature. In other words, humans understand the image qualitatively but machine recognizes the image quantitatively. Therefore, we proposed to use semantic multimodality ontology with relevance feedback approach to improve the retrieval efficiency and subsequently measure the effectiveness of retrieval such as precision and recall.
