Realistic Character Animation Research Articles

Efficient motion capture data recovery via relationship-aggregated graph network and temporal pattern reasoning.

Human motion capture (mocap) data is of crucial importance to the realistic character animation, and the missing optical marker problem caused by marker falling off or occlusions often limit its performance in real-world applications. Although great progress has been made in mocap data recovery, it is still a challenging task primarily due to the articulated complexity and long-term dependencies in movements. To tackle these concerns, this paper proposes an efficient mocap data recovery approach by using Relationship-aggregated Graph Network and Temporal Pattern Reasoning (RGN-TPR). The RGN is comprised of two tailored graph encoders, local graph encoder (LGE) and global graph encoder (GGE). By dividing the human skeletal structure into several parts, LGE encodes the high-level semantic node features and their semantic relationships in each local part, while the GGE aggregates the structural relationships between different parts for whole skeletal data representation. Further, TPR utilizes self-attention mechanism to exploit the intra-frame interactions, and employs temporal transformer to capture long-term dependencies, whereby the discriminative spatio-temporal features can be reasonably obtained for efficient motion recovery. Extensive experiments tested on public datasets qualitatively and quantitatively verify the superiorities of the proposed learning framework for mocap data recovery, and show its improved performance with the state-of-the-arts.

Efficient and realistic character animation through analytical physics-based skin deformation

Physics-based skin deformation methods can greatly improve the realism of character animation, but require non-trivial training, intensive manual intervention, and heavy numerical calculations. Due to these limitations, it is generally time-consuming to implement them, and difficult to achieve a high runtime efficiency. In order to tackle the above limitations caused by numerical calculations of physics-based skin deformation, we propose a simple and efficient analytical approach for physics-based skin deformations. Specifically, we (1) employ Fourier series to convert 3D mesh models into continuous parametric representations through a conversion algorithm, which largely reduces data size and computing time but still keeps high realism, (2) introduce a partial differential equation (PDE)-based skin deformation model and successfully obtain the first analytical solution to physics-based skin deformations which overcomes the limitations of numerical calculations. Our approach is easy to use, highly efficient, and capable to create physically realistic skin deformations.

Performance Driven-biped Control for Animated Human Model with Motion Synthesis Data

Games and 3D movies are mostly supported by realistic character animation performance. The behaviour of the humanoid character it is depend on the motion data itself. Therefore the complexity of character movement will determine the realism of their behaviour. Generally, Motion Capture device will provide the raw data that previously recorded from human/actor movement. However there are some problems remains challenges such as controller, physic effect, or motion combination. Our proposed approach will read the data from motion capture device then transformed into realistic behaviour in virtual environment. However, there are few difficulties on realizing this idea, such as user objective and the appropriate behaviour of virtual human. Therefore, we solve this issue by providing the biped control to overcome the complexity of motion synthesis data when it will be applied into character animation. The controller is capable to perform motion blending with inverse and forward kinematics, as a result it able to generate the realistic behaviour along with user intention. There is three main behaviour walking, steady and jogging that has value 0-100. As a result of experiment, the biped interface control is able to read data from motion capture then load and control the virtual human by manipulating the joint forces power in every movement of the characters. As future works, the external physical forces can be added as additional forces in humanoid model to provide certain effect such as: falling down, jumping or kicking and punching to generate realistic motion synthesis.

Tworzenie realistycznych animacji postaci 3D na podstawie danych pozyskanych z systemu wizyjnego dla poprawienia jakości filmów edukacyjnych dotyczących tematyki bezpieczeństwa

Tworzenie realistycznych animacji postaci 3D na podstawie danych pozyskanych z systemu wizyjnego dla poprawienia jakości filmów edukacyjnych dotyczących tematyki bezpieczeństwa

Hierarchical block-based incomplete human mocap data recovery using adaptive nonnegative matrix factorization

Human motion capture (mocap) data has been widely utilized for realistic character animation, and the missing marker problem caused by occlusions or a marker falling off often results in an incomplete collection. In this paper, we present a hierarchical block-based incomplete human mocap data recovery approach by using adaptive nonnegative matrix factorization, which mainly consists of two layers: interior layer and exterior layer. In the interior layer, we first decompose the underling human skeleton model into five blocks and represent the whole human mocap data in terms of the block-based sub-chain motion clips, in which the moving trajectories of each sub-chain motion clip always share the approximately low-rank property. Then, an adaptive nonnegative matrix factorization method aiming at exploiting the low-rank structure and the nonnegativity constraint is presented to restore each incomplete sub-chain motion clip individually. In the exterior layer, we integrate the recovered sub-chain motion clips and further utilize the known entries within the raw mocap data to refine the corresponding restored data of same positions, whereby the whole incomplete human mocap data can be well recovered. Without any user assistance and the training priors, the experimental results have shown the reliable recovering performance in comparison with the state-of-the-art competing approaches.

Hybrid motion graph for character motion synthesis

ObjectiveThis paper proposes a novel framework of Hybrid Motion Graph (HMG) for creating character animations, which enhances the graph-based structural control by motion field representations for efficient motion synthesis of diverse and interactive character animations. MethodsIn HMG framework, the motion template of each class is automatically derived from the training motions for capturing the general spatio-temporal characteristics of an entire motion class. Typical motion field for each class is then constructed. The smooth transitions among motion classes are then generated by interpolating the related motion templates with spacetime constraints. Finally, a hybrid motion graph is built by integrating the separate motion fields for each motion class into the global structural control of motion graph through smooth transition. ResultsIn motion synthesis stage, a character may freely ‘switch’ among different motion classes in the hybrid motion graph via smooth transitions between motion templates and ‘flow’ within each class through the continuous space of motion field with agile and the continuous control process. ConclusionExperimental results show that our framework realizes the fast connectivity among different motion classes and high responsiveness and interactivity for creating realistic character animation of rich behaviors with limited motion data and computational resources.

A script engine for realistic human motion generation

This paper proposes a framework of script engine for realistic character animation synthesis based on MoCap database. Given user-defined motion script sequence which describe the motion properties, the script engine decomposes them into sequential commands for retrieving relevant motion clips from the MoCap database and generate final animations. Furthermore, users can define their own motion elements table and scripts format suitable for various MoCap data sets. The experimental results show that this script engine framework can produce robust results and be used as a human motion engine in various applications, such as computer games, movies, sports simulation and virtual reality.

A Statistical Model of Human Pose and Body Shape

AbstractGeneration and animation of realistic humans is an essential part of many projects in today's media industry. Especially, the games and special effects industry heavily depend on realistic human animation. In this work a unified model that describes both, human pose and body shape is introduced which allows us to accurately model muscle deformations not only as a function of pose but also dependent on the physique of the subject. Coupled with the model's ability to generate arbitrary human body shapes, it severely simplifies the generation of highly realistic character animations. A learning based approach is trained on approximately 550 full body 3D laser scans taken of 114 subjects. Scan registration is performed using a non‐rigid deformation technique. Then, a rotation invariant encoding of the acquired exemplars permits the computation of a statistical model that simultaneously encodes pose and body shape. Finally, morphing or generating meshes according to several constraints simultaneously can be achieved by training semantically meaningful regressors.

Perceptual metrics for character animation

Motion capture data and techniques for blending, editing, and sequencing that data can produce rich, realistic character animation; however, the output of these motion processing techniques sometimes appears unnatural. For example, the motion may violate physical laws or reflect unreasonable forces from the character or the environment. While problems such as these can be fixed, doing so is not yet feasible in real time environments. We are interested in developing ways to estimate perceived error in animated human motion so that the output quality of motion processing techniques can be better controlled to meet user goals.This paper presents results of a study of user sensitivity to errors in animated human motion. Errors were systematically added to human jumping motion, and the ability of subjects to detect these errors was measured. We found that users were able to detect motion with errors, and noted some interesting trends: errors in horizontal velocity were easier to detect than errors in vertical velocity, and added accelerations were easier to detect than added decelerations. On the basis of our results, we propose a perceptually based metric for measuring errors in ballistic human motion.

A solid volumetric deformable muscle model for computer animation using weighted residual method

One of the most important yet expensive tasks in realistic character animation is to represent the body deformation, i.e. the deformation of the muscles. Because of the complexity of the geometry and deformable behaviours of such muscle tissues, three-dimensional (3D) finite element analysis (FEA) is usually employed. FEA however, requires enormous computing power, and often is impractical for computer animation where computational cost needs to be balanced with modelling accuracy. In this sense, it differs from engineering analysis where accuracy is often paramount for many applications, such as in aviation industry. Proposed in this paper is a 3D solid volumetric deformable muscle model which is almost as accurate as the FEA, but requires only a fraction of the computing cost. In this model, the muscle tissues are assumed to be elastic bodies with isotropic behaviour. The governing equations are derived and a weighted residual method is introduced to solve these equations. Suitable trial functions are proposed which satisfy the deformation compatibility and equilibrium equations. Using the least squares technique, the residual values on the boundaries are formulated and minimised to determine the unknown constants in the trial functions. In order to validate our model, the numerical results produced by our model are compared with the finite element computation and it is found that the two results closely agree with each other.

Controlling physics in realistic character animation

Controlling physics in realistic character animation