Research On Key Technologies Of Content-Basedvideoemotion Analysis

With the continuous development of multimedia technology, digital video in recent years has gradually become an important way of information documentation and dissemination. The traditional manual annotation methods can not adapt to the rapid growth of video data classification and retrieval needs. In this case, people began to study the content-based video semantic analysis technology, through the establishment of the video low-level physical characteristics of space and high-level semantic space mapping, obtain the video semantic content description of video automatically. Extracted from the video content can reflect the human subjective concept of semantic information, how to describe the emotion expressed by the video information are becoming the new focus of the video content analysis, is one of the important research directions in the video content analysis.Content-based video emotional semantic analysiscombines a number of technologies, the major critical technologies including emotional semantic modeling, cognitive level semantic analysis, emotional layer semantic analysis and the emotional content-based video retrieval technique, this paperthese aspects in-depth analysis and research, the work done and the major innovations are as follows:(1) Design a new emotional space model.By the basic theory and the video characteristics of emotional information, usingemotional histogram way to express the emotions generated video emotional space, the model can effectively carry out the description of the emotional type and emotionalintensity of expression.(2) Design ontology-based video emotionalsemantic analysis framework.By the basic law of semantic content presented in the video and the research results of the ontology, this paper deifned the concept of the video ontology, and built a function system and reasoning relations. According to this framework,this paperdesign a video emotional semantic analysis of the overall process.(3) Proposed a cognitive level semantic analysis algorithms, cognitive level semantic analysis is an important step of the emotion recognition, the cognitive level semantic analysis include video object segmentation algorithm, the classification algorithm of video object recognition and object tracking algorithm, the study proposed a set of improved algorithms and thus a comparative analysis.(4) Proposed the concept of the video "emotional semantic object". This paper analyzes the human visual system (HVS) attention mechanisms, and to draw attention and generate emotional response to the main object identification to determine the video. Artificial neural network technology into the video field of sentiment analysis, and constructed a set of semantics from a cognitive layer to the emotional layer of semantic mapping model.(5) Designed a set of emotional content-based video retrieval system. Using object-oriented technology and relational database technology, the semantic content of the video analysis to obtain the emotional layer structured way to store and organize, to provide an effective user interface search video affective content.