Görüntü veritabanı sistemi için önemli nesnelerin otomatik olarak bulunması

Recently, the increase in the amount of multimedia data has unleashed the devel-opment of storage techniques. Multimedia databases is one of the most popularof these techniques because of its scalability and ability to be queried by themedia features. One downside of these databases is the necessity for processingof the media for feature extraction prior to storage and querying. Ever growingpile of media makes this processing harder to be completed manually. This is thecase with BilVideo Video Database System, as well. Improvements on computervision techniques for object detection and tracking have made automation of thistedious manual task possible. In this thesis, we propose a tool for the automaticdetection of objects of interest and deriving spatio-temporal relations betweenthem in video frames. The proposed framework covers the scalable architecturefor video processing and the stages for cut detection, object detection and track-ing. We use color histograms for cut detection. Based on detected shots, thesystem detects salient objects in the scene, by making use of color regions andcamera focus estimation. Then, the detected objects are tracked based on theirlocation, shape and estimated speed.Keywords: Video Databases, Video Object Detection, Object Tracking, CameraFocus Estimation.i