Automated Shot Boundary Detection in VIRS DJ Park Computer Science Department The University of Iowa
Introduction VIRS? n Visual Information Retrieval System n Works on multimedia information domain such as video, sound, etc n Aims for content-based analysis, retrieval and presentation n Find all videos that Iowa football team plays n
Building VIRS Shot Boundary Detection n Scene Boundary Detection n Key Frame Generation n Pattern Recognition n Indexing with Meta Data n Interface and Querying Method n
Shot Boundary Detection A sequence of frames captured by a single camera in a single continuous action n Most basic unit of video data n A scene is a logical group of shots into a semantic unit n Many different styles and techniques ( cut, fade, wipe, dissolve ) n How to detect among these styles? n What about camera/object movements? n
General Approach Frame to Frame Comparison n Take 2 subsequent frames and measure their differences n How to define these differences discriminates each method n Pixel Difference n Color Histogram Difference n Edge Difference n
Pixel Difference One of the earliest methods n How many pixels are different between subsequent frames? n Problems? n Improvement – using blocks instead of pixels n (X, Y) Frame t+1
Color Histogram Difference Much more accurate than simple pixel-wise approach n Compare the number of occurrences for each color between subsequent frames n Problems? n Frame t Histogram t
Edge Difference n n No color information is needed Compare entering edge pixels and exiting edge pixels Better than other methods in detecting fade, wipe, dissolve Problems?
Future Works Evaluation and Combination of existing methods n Relating shot boundary information with scene boundary – how to group shots into semantic unit? n Using sound information for shot ( or scene ) boundary detection? n Using object-oriented approach or 3 -D image modeling technique? n
Questions?
Скачать презентацию Automated Shot Boundary Detection in VIRS DJ Park
c5b44665fd7df8858e17f8ca12ea0c0b.ppt