Computer Facial Animation: A Review

Abstract

Abstract-Computer facial animation is not a new endeavour as it had been introduced since 1970s. However, animating human face still presents interesting challenges because of its familiarity as the face is the part used to recognize individuals. Facial modelling and facial animation are important in developing realistic computer facial animation. Both modelling and animation is dependent to drive the animation. This paper reviews several geometric-based modelling (shape interpolation, parameterization and muscle-based animation) and data-driven animation (image-based techniques speech-driven techniques and performance-driven animation) techniques used in computer graphics and vision for facial animation. The main concepts and problems for each technique are highlighted in the paper.Index Terms-Computer graphics, data-driven animation, facial animation, geometric-based modelling. I. INTRODUCTIONComputer facial animation is experiencing a continuous and rapid growth since the pioneering work of Frederick I. Parke [1] in 1972. This is partly due to the increasing demand of virtual characters or avatars in the field of gaming, filming, human computer interaction and human machine communication.The human face is one of the channels in expressing the affective states. It has complex but flexible three-dimensional (3D) surfaces. To present the human face onto a computer system is a challenging task due to several goals to be achieved. According to Deng and Noh [2], the main goal in facial modelling and animation research is to develop a high adaptability system that creates realistic animation in real time mode where decreasing manual handling process. Here, high adaptability refers to the system that is easily adapted to any individuals" face.Various approaches have been proposed by the research community to improve the performance of facial animation in different aspects. Some claims that a good facial modelling and animation involving how to lip synchronization and how to use the eyes, brows and lids for expression [3]. Others believed that a facial model should include other attributes such as surface colours and textures [4].The focus of this paper is to present a review of the past and recent advancements in facial animation research. Besides, an overview of the facial animation framework is discussed to give a picture of where the modelling and animation techniques take place. The goal of this review is to improve the current computer facial animation system with the knowledge of the current advancement in this technology. II. A GENERAL FACIAL ANIMATION FRAMEWORKAnimating facial action is a tedious and complicated task. The construction of facial animation proceeds in several stages and is shown in Fig. 1. Each facial animation starts from the head model creation. The facial geometry is created from the input data either captured from images or motions. The input data are pre-processed and filtered, only the corresponding data is used to create the basis of morphable meshes. Next, geometric-based modelling is a...

Synthesizing facial expressions for signing avatars using MPEG4 feature points

“…near muscle (2) nts in its influence f attachment. The muscular action is ant representing the only two instances muscle around each es the mouth.…”

“…However, it is important to note here that the generation of a significant range of highly detailed expressions usually implies the creation of large libraries of blend shapes which can be very time-consuming. Moreover, if the topology of the model needs to be changed, all the shapes must be redone [2].…”

A survey on the pipeline evolution of facial capture and tracking for digital humans

“…Later in 2013, Agianpuye and Minoi [2] published a survey that focuses more on the different facial animation approaches used in the literature rather than the 3D reconstruction of the face, but it includes a wide selection of facial expression synthesis methods, including other statistical and learning-based methods. In the same year, Ping et al [90] presented a survey that reviews geometric-based modeling for face representations and data-driven animation techniques for facial animation. These surveys talk about some of the facial animation approaches mentioned in this work, as well as the same facial codifications used for that task, however, this survey complements both works by mentioning the various available solutions that offer a complete reconstruction and animation framework with the use of one or several of the discussed animation approaches.…”

Facial Expression Modeling and Synthesis for Patient Simulator Systems: Past, Present, and Future

“…Facial expression synthesis and animation (FSA) refers to techniques used to animate dynamic expressions on the faces of avatars or robots using previously developed face models. FSA techniques provide the facial movement vocabulary that maps the developed model of AU movements and densities into the mesh topology of the social robot or avatar heads [148]. Using this technique makes the simulated face able to display AU Rigs a skeletal model to automatically associate each bone and joint into various parts of the embodiment's face and animates it using skeletal motion data.…”