A new global motion estimation algorithm and its application to retrieval in sports events

Abstract

In this paper, we propose a novel global motion estimation technique based on weighted gradient and Displaced Frame Difference (DFD) associated with Wiener estimation. Then, we apply this technique to parse events of a high level of understanding in a cricket game. A user oriented analysis of the game then reveals a distinct connection between the global motion and specific events. By estimating global motion and analysing the temporal evolution of the estimated motion parameters, we present an effective process for the extraction of cricket events, leading to a succes rate of 88.9%.

Browsing sports video: trends in sports-related indexing and retrieval work

“…It therefore stands to reason that the motion of the camera contains semantic information. In cricket, Kokaram et al [13] show that the motion of the camera can be used to detect when a play is about to start, the duration of the play, and the direction of the ball after it is hit. Action in soccer [14] can also be characterized in this manner.…”

“…In [13], a shot in cricket was detected by a strong horizontal pan in the camera motion, while in [12] and [16], the motion of objects themselves is used to classify an entire play sequence in tennis and snooker. Statistical modeling of motion content has also been proposed for classification of sequences in different sport videos, such as skating and athletics [15].…”

Broadcast Cricket Highlight Treatment using a Quadrant Based Method for Global Motion

“…[13] analysed pre-classified long-view shots for soccer using Global Motion Estimation to identify the team with the advantage during the shot. Kokaram and Delacourt [14] provided a new algorithm for global motion features and applied it to characterize the main views associated with cricket sequences.…”

Exploiting temporal discontinuities for event detection and manipulation in video streams

“…Shot changes are in turn the building block of semantic analysis [15,3]. Camera motion is another such feature, being useful for identifying shot changes [2] as well as semantic events in cricket [9]. Object motion is clearly another universally useful and semantically relevant feature, although much more difficult to extract [14,4].…”