Cyclic Motion Analysis Using the Period Trace

TLDR: Cyclic motion analysis techniques have broad applicability because many real-world motions repeat, e.g., a heart beating, an athlete running, and a wheel rotating, so they can be formally described and interpreted regardless of the particular object or scene that is moving.

Abstract: Non-rigid motion analysis is complicated by a lack of general purpose rules or constraints governing how an object or scene evolves over time. In order to design practical algorithms, much of the work to date has focused on object-model-based techniques, such as interpretation of facial expressions, detection of human locomotion, cardiac image analysis, and gesture recognition. In contrast, repeating motions all share common temporal features that can be formally described and interpreted regardless of the particular object or scene that is moving. For instance, it is not necessary to recognize the runner or the motion in Fig. 1 in order to determine his stride frequency. Because many real-world motions repeat, e.g., a heart beating, an athlete running, and a wheel rotating, cyclic motion analysis techniques have broad applicability.