Research Of Image Emotion Ontology Connecting With FCA

The article's research background is the combination of image emotion and ontology and the construction and renovation are the start points of the whole research. People's real sense of the picture produced from many factors, not only including the objective elements of the image itself but also a particular feeling, mood and other emotional subjective experience inspired by the image. Many scholars uses the key words which generated by the clients, such as beautiful, warm, cool, as the search keywords of corresponding image to study how to connect the subjective feeling with image. As the description vector in the knowledge fields, Ontology can simulate human emotional recognition scene through the various relations in its structure, in order to effectively solve the emotional image problems. It is very important to study different knowledge of the ontology construction to clearly build the image emotional ontology with the artificial construction or the automatic one. This article summarizes a variety of language and tools for learning ontology. These tools can help us in every structure and every point of the image ontology learning.In order to achieve a semi-automatic or automatic way to build the image emotional ontology, we must have a reasonable and efficient ontology construction and learning method. However, the modern domestic and international researches are numerous, but there is no universal and effective solution for the image emotional ontology construction. As the two formal methods, Formal concept analysis and ontology have few differences. They are similar in the structure and function of the concept lattice, which is the core structure of them. They both stress on concept's inter-subjective consensus and the model form description. We can find their contact and connection way to built the ontology in a more rational and efficient form, for semi-automatic or automatic machinery learning.The article further combines the formal concept analysis to the ontology learning and proposes a new auxiliary idea for the emotional ontology learning. It has the following four parts:First, the article describes the using of Image Emotion ontology contributed by MPEG-7 framework descriptors and other ontology learning tools and construction methods. These tools can help us in every structure and every point of the image ontology learning.Second, it introduces the fuzzy formal concept analysis and the similarities between the formal concept analysis and the Image Emotion Ontology Construction. Through taking the core of Formal Concept Analysis theory, Concept Lattice, into the fuzzy formal context which uses the image underlying characteristics as the extension and the emotion distribution of VAD mood as the content, it gives a fuzzy formal concept lattice theory based on the image emotion ontology, as well as proposes a algorithm research on constructing and maintaining the appropriate concept lattice according to the fuzzy formal context. This concept lattice which appeared by the appropriate formal background can create Image Emotion ontology through the structure similarities between Lattice and ontology and the ontology learning method.Third, it compares the different features of various construction algorithm among the concept lattices and fuzzy ones. It has proved that it has stability, time and space superiority compared to the other representative ones before by experiment. It also proves that it is feasible to generate the fuzzy concept lattice and further apply to the image feeling ontology learning by the corresponding upper and lower thresholds with fuzzy formal context.Fourth, it proposes further research directions and provides a support to other studies in the same topic group.