This paper presents an improved frame-to-frame (F-2-F) compressedudvideo matching technique based on local features extracted from reduced sizeudimages, in contrast with previous F-2-F techniques that utilized global fea-udtures extracted from full size frames. The revised technique addresses bothudaccuracy and computational cost issues of the traditional F-2-F approach. Ac-udcuracy is improved through using local features, while computational cost issueudis addressed through extracting those local features from reduced size images.udFor compressed videos, the DC-image sequence, without full decompression, isudused. Utilizing such small size images (DC-images) as a base for the proposedudwork is important, as it pushes the traditional F-2-F from o�-line to real-timeudoperational mode. The proposed technique involves addressing an importantudproblem: namely the extraction of enough local features from such a small sizeudimages to achieve robust matching. The relevant arguments and supportingudevidences for the proposed technique are presented. Experimental results andudevaluation, on multiple challenging datasets, show considerable computationaludtime improvements for the proposed technique accompanied by a comparableudor higher accuracy than state-of-the-art related techniques.
展开▼
