Journal article Open Access

Deep Motifs and Motion Signatures

Aristidou Andreas; Cohen-Or Daniel; Hodgins Jessica K; Chrysanthou Yiorgos; Shamir Ariel


JSON-LD (schema.org) Export

{
  "inLanguage": {
    "alternateName": "eng", 
    "@type": "Language", 
    "name": "English"
  }, 
  "description": "<p>Many analysis tasks for human motion rely on high-level similarity between sequences of motions, that are not an exact matches in joint angles, timing, or ordering of actions. Even the same movements performed by the same person can vary in duration and speed. Similar motions are characterized by similar sets of actions that appear frequently. In this paper we introduce motion motifs and motion signatures that are a succinct but descriptive representation of motion sequences. We first break the motion sequences to short-term movements called motion words, and then cluster the words in a high-dimensional feature space to find motifs. Hence, motifs are words that are both common and descriptive, and their distribution represents the motion sequence. To cluster words and find motifs, the challenge is to define an effective feature space, where the distances among motion words are semantically meaningful, and where variations in speed and duration are handled. To this end, we use a deep neural network to embed the motion&nbsp;words into feature space using a triplet loss function. To define a signature, we choose a finite set of motion-motifs, creating a bag-of-motifs representation for the sequence. Motion signatures are agnostic to movement order, speed or duration variations, and can distinguish fine-grained differences between motions of the same class. We illustrate examples of characterizing motion sequences by motifs, and for the use of motion signatures in anumber of applications.</p>", 
  "license": "http://creativecommons.org/licenses/by-nc-nd/4.0/legalcode", 
  "creator": [
    {
      "affiliation": "The Interdisciplinary Centre", 
      "@type": "Person", 
      "name": "Aristidou Andreas"
    }, 
    {
      "affiliation": "Tel-Aviv University", 
      "@type": "Person", 
      "name": "Cohen-Or Daniel"
    }, 
    {
      "affiliation": "Carnegie Mellon University", 
      "@type": "Person", 
      "name": "Hodgins Jessica K"
    }, 
    {
      "affiliation": "Research Centre on Interactive Media Smart Systems and Emerging Technologies", 
      "@type": "Person", 
      "name": "Chrysanthou Yiorgos"
    }, 
    {
      "affiliation": "The Interdisciplinary Centre", 
      "@type": "Person", 
      "name": "Shamir Ariel"
    }
  ], 
  "headline": "Deep Motifs and Motion Signatures", 
  "image": "https://zenodo.org/static/img/logos/zenodo-gradient-round.svg", 
  "datePublished": "2018-11-01", 
  "url": "https://zenodo.org/record/2658798", 
  "version": "Published", 
  "keywords": [
    "Motion capture", 
    "Motion processing", 
    "Animation", 
    "Motion Word", 
    "Motif", 
    "Motion Signature", 
    "Convolutional Network,", 
    "Triplet Loss"
  ], 
  "@context": "https://schema.org/", 
  "identifier": "https://doi.org/10.1145/3272127.3275038", 
  "@id": "https://doi.org/10.1145/3272127.3275038", 
  "@type": "ScholarlyArticle", 
  "name": "Deep Motifs and Motion Signatures"
}
70
40
views
downloads
Views 70
Downloads 40
Data volume 1.0 GB
Unique views 56
Unique downloads 39

Share

Cite as