Efficient motion retrieval in large motion databases

Abstract

There has been a recent paradigm shift in the computer animation industry with an increasing use of prerecorded motion for animating virtual characters. A fundamental requirement to using motion capture data is an efficient method for indexing and retrieving motions. In this paper, we propose a flexible, efficient method for searching arbitrarily complex motions in large motion databases. Motions are encoded using keys which represent a wide array of structural, geometric and, dynamic features of human motion. Keys provide a representative search space for indexing motions and users can specify sequences of key values as well as multiple combination of key sequences to search for complex motions. We use a trie-based data structure to provide an efficient mapping from key sequences to motions. The search times (even on a single CPU) are very fast, opening the possibility of using large motion data sets in real-time applications.

Point-based Medialness for Movement Computing

“…The Line of Action which can be interpreted both as a simplification of medialness (a single medial curve) and an extension (applicable to more than one form at a time) can potentially be represented on the basis of our medial point set by selecting and interpolating a subset. Such an interpolation can also be used to recover a skeletal graph or sketch which can feed in recently proposed information retrieval methods applied to video archives of dance footage [14,13]. We are also exploring the relation between drawing techniques using primitives (like disks) to create a scaffolding for a final drawn rendering of a moving form (Fig.…”

“…Even today 3D animators in games and films create key-frames, but now the in-betweening is made with software. Lately there has been a rise of interest in the notion of key-framing for notating movement as it cuts out extra data [14,13].…”

Laban movement analysis and hidden Markov models for dynamic 3D gesture recognition

“…The processed results are compared to specialists' annotation, and the matching occurs only for certain qualities. In [26], expressive features aiming at quantifying Body, Effort and Shape LMA qualities (cf. Section 3.1) are defined locally (i.e., for each frame of a gesture), in order to index various gestural contents with local motion "keys".…”

Untitled

“…Initially, the objective in this domain was not to access perceptual processes only but also to create complementary strategies to organize and classify large databases on motion that included more subtle aspects of expressions relating to emotion. A first attempt by Wakayama, Okajima, Takano, and Okada (2010), then Okajima, Wakayama, and Okada (2012), showed that motion retrieval can benefit from the use of a subset of LMA dimensions, especially when searching for data (indications, instances) on body movement in large research databases (Kapadia, Chiang, Thomas, Badler, & Kider, 2013). Recently, Aristidou, Charalambous, and Chrysanthou (2015) inspected the similarities among various emotional states classified according to the arousal and valence of Russell's (1980) circumplex model and a subset of features that encode stylistic characteristics of motion based on the LMA.…”