Characteristics of video data for signal analysis

Video data is complex, unstructured and voluminous unlike conventional text data. For providing effective retrieval facilities, video data needs to be stored in a database with an appropriate data structure. Hence, identifying the required information units and formulating a data structure have become the fundamental tasks. Data mining methods analyze data for extracting patterns. In this paper, we elaborate the characteristics of video data which must play a significant role in the development of the data mining algorithm. We have discussed the characteristics of video data under three classes, namely, dominant, native and assigned which have influenced us in arriving at a scheme for analysis.