H. Koumaras National Ctr of Scientific Research "Demokritos" (Greece) G. Gardikis Univ. of the Aegean (Greece) G. Xilouris National Ctr of Scientific Research "Demokritos" (Greece) E. Pallis Technological Education Institute-Crete (Greece) A. Kourtis National Ctr of Scientific Research "Demokritos" (Greece)
Create and organize publications into your own personal collections/lists
Easily search saved publications across your mulitple lists
Share your collections with friends, coworkers, or anyone that might be interested in the same research
To take advantage of My Library, sign in now.
Automatic shot boundary detection is a field, where many techniques and methods have been proposed and have claimed to perform reliably, especially for abrupt scene cut detection. However, all the proposed methods share a common drawback: the necessity of a threshold value, which is used as a reference for detecting scene changes. The determination of the appropriate value or the dynamic reestimation of this threshold parameter remains the most challenging issue for the existing shot boundary detection algorithms. We introduce a novel method for shot boundary detection of discrete cosine transform (DCT)-based and low-bit-rate encoded clips, which exploits the perceptual blockiness effect detection on each frame without using any threshold parameter, therefore minimizing the processing demands required for algorithm implementation.