Study on Event Matching In Temporal Database Using AGT Approach

At some instant, the user may want to acquire the event as image from the temporal database. Prior to the retrieval, the system has to be trained with the data based on the events. Firstly, the video file is converted into frames with the time range with a set of frames. Each frame i.e., image is subjected to image processing techniques to get the various features of that image such as RGB values, histograms, hue, saturation, intensity, time information, etc. These extracted features are stored in the temporal database. The image retrieval is based on the events and the user has to give the query in natural language query. Based on the natural language query, some set of images are retrieved from the database and these retrieved images are used as the query images. User can select any one of the image as the query image and give for the processing. Now the query image is subjected to the image processing technique and these techniques are used to extract the various features of query image. The proposed system is provided with the query image along with its features. Thus the SQL query is formed with NL query and features of the image. Then the image that matches exactly with the requirements is fetched from the database.