Compression of Human Motion Capture Data Using Motion Pattern Indexing

Gu, Qin; Peng, Jingliang; Deng, Zhigang

doi:10.1111/j.1467-8659.2008.01309.x

Cited by 41 publications

(39 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When applying the proposed method on these MoCap files, there was no significant difference between the compression and decompression times. Table 4 compares the processing time of the proposed method with other methods [2,1,4], showing that (1) the proposed method satisfies real-time application requirements and (2) the proposed method is better than the other methods in terms of time and quality performance at the same compression ratio. The results in [2] were measured on a P4 3.4 with 3 GB of RAM as reported in the paper published in 2006.…”

Section: Methodsmentioning

confidence: 93%

“…While [1] performs well on shorter motion clips, the compression is very time consuming because of the optimization process. The method proposed in [4] requires a preprocessing phase to create a database of motion primitives making it unsuitable for realtime compression of short clips. Table 3 compares the results of compressing the motion capture files with different weight combinations given to kinematic chains.…”

Section: Methodsmentioning

confidence: 99%

“…MoCap compression can also be achieved by defining a database of motion primitives and replacing each segment of the motion with the closest match in the database [4]. The main drawback of this approach is that the database needs to be transmitted along with the compressed data and therefore the encoding is not efficient for small motion sequences.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Perceptually motivated LSPIHT for motion capture data compression

Cheng

Firouzmanesh

Basu

2015

Computers & Graphics

View full text Add to dashboard Cite

Section: Methodsmentioning

confidence: 93%

Section: Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Perceptually motivated LSPIHT for motion capture data compression

Cheng

Firouzmanesh

Basu

2015

Computers & Graphics

View full text Add to dashboard Cite

“…1, motions could be represented by a series of key-poses. Therefore, all poses in a database could be clustered, and the idea of using motion clustering indices (MCI) to represent poses is adopted by [21,24,25]. However, all numeric-based features cannot describe the logic meaning of a motion, e.g.…”

Section: Related Workmentioning

confidence: 99%

A human motion feature based on semi-supervised learning of GMM

et al. 2014

View full text Add to dashboard Cite

those motion data to create realistic motions for new characters by applying motion editing [1,2], motion synthesis [3][4][5] and motion retargeting [6] techniques has become the focus of research in the past several years. However, prior to reusing and processing the old motions, one fundamental problem of identifying and extracting similar motion clips from the database has to be solved. It is essentially a motion matching problem. The general procedure is to calculate a concise and representative feature for each motion, and then compare the similarities with all other motions in the mocap database. The efficiency and accuracy of these motion retrieval and annotation processes largely depend on the property of the used features.Most early work [7] uses textual description such as running, walking, to label existing motions in a database. It does not only involve a lot of manual work, the textual label is also too short and general to fully represent the features of each motion. Later works [8][9][10][11] use the numeric-based features and logic-based features involving the 3D coordinates of each joint in each frame. It includes too much redundant information and the 'huge' feature makes the motion matching really slow. Some recent works [12] present semantic features which better represent the essence of motions and the low dimension of features largely speeds up the motion retrieval process. In this paper, we present a new feature in this category. The work [13] has shown that a human motion clip could be described with some representative poses, which we call the 'key-pose'. Intuitively, two similar motions may share most key-poses, while the motions belonging to different motion classes may share none or only a few key-poses (as shown in Fig. 1). A good selection of key-poses can be used to represent different motion classes. The second benefit to use key-poses as a feature is, although the category of motions is infinite, the types of key-poses are relatively limited. A new category Abstract Using motion capture to create naturally looking motion sequences for virtual character animation has become a standard procedure in the games and visual effects industry. With the fast growth of motion data, the task of automatically annotating new motions is gaining an importance. In this paper, we present a novel statistic feature to represent each motion according to the pre-labeled categories of key-poses. A probabilistic model is trained with semi-supervised learning of the Gaussian mixture model (GMM). Each pose in a given motion could then be described by a feature vector of a series of probabilities by GMM. A motion feature descriptor is proposed based on the statistics of all pose features. The experimental results and comparison with existing work show that our method performs more accurately and efficiently in motion retrieval and annotation.

show abstract

“…Michalis Taptis applies logistic regression to evaluate how well the dancer's performance matches with the canonical model [5] and obtains good results, however, motion curves, which are its proposed features, are not good abstraction but rather another way of presentation of motion data. Third, methods based on statistical model include HMM [6]- [8], DBN (Dynamic Bayes Network) [9], etc. Well trained statistic model can unveil the implicit statistical relations in gestures or motion sequences, thus help eliminate spatio-temporal variations in recognition process.…”

Section: Introductionmentioning

confidence: 99%

Real-Time 3D Motion Recognition of Skeleton Animation Data Stream

Lv¹,

Xiao²

2013

IJMLC

View full text Add to dashboard Cite

Abstract-In this paper, a method for real-time 3D motion recognition based on a hierarchical recognition framework is presented. To facilitate the recognition process, motions are divided into three levels by duration and complexity. SVD (Singular Value Decomposition) is used to extract the feature vector of each motion matrix, and SVM(Support Vector Machine) is utilized to do the training and classification of the first level of motion(sub-motion). In motion recognition process, the sequence of recognized candidate sub-motions is analyzed by HMM (Hidden Markov Model) to gain certain robustness, then we recognize the second level of motions by pattern matching in this sequence. Finally a grammar-based motion synthesization is applied using motions as semantic terms to recognize the third level of motions. Experimental results show that the proposed method has high performance in sensitivity, accuracy, specialty and efficiency.

show abstract

Compression of Human Motion Capture Data Using Motion Pattern Indexing

Cited by 41 publications

References 25 publications

Perceptually motivated LSPIHT for motion capture data compression

Perceptually motivated LSPIHT for motion capture data compression

A human motion feature based on semi-supervised learning of GMM

Real-Time 3D Motion Recognition of Skeleton Animation Data Stream

Contact Info

Product

Resources

About