Image processing on encoded video sequences

Abstract

Abstract. This paper presents a novel approach to processing encoded video sequences prior to complete decoding. Scene changes are easily detected using DCT coefficients in JPEG and MPEG encoded video sequences. In addition, by analyzing the DCT coefficients, regions of interest may be isolated prior to decompression, increasing the efficiency of any subsequent image processing steps, such as edge detection. The results are currently used in a video browser and are part of an ongoing research project in creating large video databases. The procedure is detailed with several examples presented and studied in depth.

Automatic and fast temporal segmentation for personalized news consuming

“…Till now, various works have been done to detect shot boundaries automatically. For example, the frame similarity/difference can be computed by pixel differences (Zhang et al 1993;Hampapur et al 1994), statistical differences (Nam and Tewfik 2005), histogram differences (Truong et al 2000), compression differences (Arman et al 1994;Joyce and Liu 2006), edge changes (Zabih et al 1999), motion vectors (Zhang et al 1993), and mutual information between pixels (Cernekova et al 2006). After getting the differences/similarities, the decisions are made by such methods as threshold-based decision (Gao and Tang 2002), B-spline based decision (Ngo 2003 Fig.…”

Content-Based Multimedia Retrieval

“…Shot boundaries are typically found by computing an image-based distance between adjacent frames and noting when the distance exceeds a certain threshold. The distance between adjacent frames can be based on statistical properties of pixels [48], compression algorithm [7], or edge differences [94]. The most widely used method is based on histogram differences.…”

Automatic video temporal segmentation based on multiple features

“…For example, the method in (Little et al 1993) uses differences in the size of JPEG compressed frames to detect shot boundaries as a supplement to a manual indexing system. The method in Arman et al (1994) detects shot boundaries by comparing a small number of connected regions. It uses differences in the discrete cosine transform (DCT) coefficients of JPEG compressed frames as their measure of frame similarity, thus avoiding the need to decompress the frames.…”