Abstract
There are lots of technical and expert supports are needed in order to bring in the actual act of motion capture. This paper is dealing with the integral factors related to human motion capture methods. At the same time the emphasis has also been laid over the various developmental aspects related to it. This is the paper that has been structured for the understanding of the chosen topic and to se the future perspective of the same field.
Introduction
The paper is related to the application of a method to provide unequal allocation of all motion capture act and motion tracking systems, or mo cap. The paper is dealing with all these terms in order to have better perception and the comprehensive follow up for the development of the same. The functionality is to describe the process of recording all sorts of movement and on the same floor getting the provision for translating the movement onto a digital model. The basic applications are in the field of military operations. Added to this there is also the provision of using it in the fields of entertainment, sports, and above all in the perspective of medical applications for the betterment of mankind. The maximum utility of it has been detected in the process of filmmaking. Aggarwal, JK and Q Cai. (1999, pp.428-440) states that in this particular genre it is referred to as medium for recording actions of human actors. As the whole movements and all the matters related to the actions get connected, there comes in the graphical representation of it. All these information are then synchronised together to animate them in the form of digital character models in 3D animation. In this act there are many distinctive participations are made to be a mark of it. The depiction or the graphical representation of face, fingers and expressions, gets transformed into digital coding and decoding. This is the act that has been termed as performance capture.
The purpose of this paper is to explore this world of capturing motion in term of the motion and the geometry that involves in the whole act. It is all very much a part of my research. My attempt is to bring in the factors related to its development and growing popularity. There is also an attempt made to bring in the major development and application in these fields. There is no doubt that by capturing motion in such a way; that the experts have brought in some of the great futuristic perspectives into the scene. There is also the need for further research in this field and these are all related to the upcoming technological and the motion picture capturing profession.
The Problem
There are certain disadvantages or as we can say, some amount of problems that are been faced in the series of making of videos that captures human motion. There were many specific hardware and many special programs are needed to make it a possible act. The process of data collection needs much sophisticated technology. Added to all these the cost of the software and all the necessary equipments, cannot be considered as affordable for small and limited productions. The proceedings related to the capturing of motion needs lots of professional expertise. There is hardly an instance that can be considered as the weakest knot. The motion capture system needs very specific requirements for the purpose to have the facility to get operated. This is a functionality that depend camera field of view. There are the critical scenes when the problems get into the scene. These are the hurdles that are very much related to the whole shooting process. There are many kinds of dominance that are preoccupied by the means of reshooting a scene rather than trying a periphery that is a part of the whole act of manipulating the data. There are very few systems that can be used to allow real time viewing of the collected and synchronized data to decide.
From geometrical point of view the act of capturing motion for quadruped characters is very much the act that needs more expertise. In certain circumstances the results that are derived are very much limited. There is the limitation related to the performance delivered and that is the reason that brings in captured mass and the entire volume without any additional or supplemented editing of the data. It is a clear understand that movements are free from any kinds of laws and as they are calculated under digital coding the geometrical existence get well sorted by it. There is also the adoption of traditional animation techniques. The basics and the primary functionality of these are very much added for the purpose of emphasizing on anticipation. This further gets followed through the means of secondary motion. There is no doubt for the manipulation of the actual shape of the particular human character. However there still remains the need for the perfect specific crafting in it. The manipulation of the shapes of characters with squash and stretch animation procedure is considered to be a hurdle. This gets into the nerve when the computer model has different proportions of depicting the collected data. It is here that the capture subject artifacts have got the possibilities to occur. This can be well scrutinized in terms of cartoon character. As a cartoon character can be depicted with large over size hand intersecting with any other body part. However a human actor brings these body parts too close to his body. The real life performance can get resembled with the graphical structure, but under no circumstance it will be having the life in it. This is one of the main problems that make the human graphical capture look more life like but less full of life.
Personal Interpretation
According to my idea the concept has grown much bigger than what it was earlier. Earlier there were many hurdles faced by the motion capture designer. But the craftsmanship has got more sophisticated and there are lots of provisions to make it a successful creation. It is true that these graphical representations are like but you can very easily bring in the difference of being lifeless. This is one of the biggest hurdles that it has got. However the latest technology has come up with lots of technological and added training for the same. The procedure is very particularly designed and the scopes are all made very sophisticated. In the process of creating motion capture there are the sessions of movements that comprises of one or more actors to act on a particular script.
On the contrary to all these troubles I personally consider that the implication of Mo cap can very distinctively offer several advantages to the act of over all traditional computer animation system under a 3D model. Some of the basic advantages are formulated for the rapid and real time results. These can be well obtained in a very short notice without wasting much time. As for the matter of affordability the implication of Mo cap in the field of entertainment applications very strongly reduces the costs of keyframe-based animation and makes it much cheaper than what it was earlier. As a matter of fact the amount of work hardly makes a difference with the complexity or length of the performance. This is very much a part of same degree when the uses of traditional techniques were thought to be better. There are various kinds of experiments and testing can be done in this respect. There are lots of variations in the presentation and the delivery provision of all kinds of styles in it.
Under the disadvantage perception the complexity gets counted to be a major drawback. However I have discovered that through proper training and exclusive interpretive interventions the complex movement and realistic physical interactions can be well knitted and captured. In case of secondary motions, the role played by weight and exchange of forces is very vital and it leads to the initiation of recreating the structure in motion in a physically accurate manner. Under modern application the basic interest is to have the act of capturing movement through an analytical survey of human motion. It has got the provision for the recovery of the motion parameters from geometrical derivations of the human body and each of the actions related to the human body parts. The possibilities can be well notified in terms of the actions related to the human hand. As against the earlier versions of interpreting actions and their depiction for the purpose of recovering constrained motions like those of walking and running; the modern approach believes more in creating the graphics as per the actions needed. Recovering unconstrained motion is a very latest kind of adoption and has been discovered to be very much authentically significant in this act of capturing human motion. As there are very large numbers of degrees of freedom, there is also equal amount of problems for sure. These problems are basically while dealing with the smaller sizes of some body parts.
Shopov, A. et al. (2000, pp. 103-110) focuses that there are the instances that have got the ambiguous consequences in case of some of the human motions these are all related to categories of self-occlusions, geometrical diversification and etc. The whole procedure of Human motion capture has wider ranges of applications. It is the out and out proceedings that include all kinds of themes like those of human monitoring, gesture analysis, surveillance, computer animation, motion recognition, and lots more of these categories. The measures include the categorical analysis of modelling, tracking and understanding of each of the human motion that gets its foundation and basis over the video sequences.
There are many researches done for the purpose of increasing importance, with the applications in biomechanics, medicine, sports, sciences, animation, surveillance, and many of these kinds. The attempt is to have better interpretation and enough dominance over this particularised section of capturing human motion. The analytical visions shows that the progress in human motion analysis depends on research in computer graphics and the sections related to computer vision and biomechanics. The intimacy of computer vision and biomechanics with human motion analysis needs a full fledged an interaction of computer graphics with computer vision. The application benefits from an understanding of biomechanic constraints.
Motion and Geometry
In order to get the whole human motion there is a huge interpretation done over all the actions and the data gets collected. All these actions get sampled first. Then there are the sessions that make these actions get modified and re-enacted for a perfect pose. These are done many times within a second. Till the graphics get all right. The techniques are all well polished and are up to date. It has been further discovered that ILM use images for 2D motion capture. After capturing it they project it into 3D motion capture. Green, RD, L Guan, and JA Burne, (2000) demonstrates that the whole act gets record records only when the dramatically collected data of particularized movements of the actor gets synchronized. In this process the actor’s visual appearance is not of much importance. It is his body that gets recorded and all his movements are collected for perfect graphical and digital intervention. It is here that the animation data gets the coverage of mapping to a 3D model. In this procedure the 3D model performs the same actions that are being collected from the live actor. All the movements and the positions are copied and are represented just like the live actor did. This was earlier done through the depiction of some comparable devices. However Sparks, C. et al. (2005) clarifies that under modern technology there is a drastic replacement done in the older technique of rotoscope. It is here that the application of the sight gets the representation through visual appearance of the motion that has been collected for the purpose. The collected data is from activities of a model actor as he gets filmed after the finishing of the whole recording process the film is very particularly used as a guide in the formation of motion capture. This is a long process that gets interpreted frame by frame motion and gets structured through hand-drawn animated character.
Müller, M., Röder, T., Clausen, M. (2005, pp. 677- 685) declares that in the whole process the most important act is done by the camera. The significant camera movements under geometrical perception give the provision to capture motion. On the other hand the virtual camera in the scene is very much into the activities of getting pan, tilt, or at times without notice gives dolly around the stage. The mechanism gets driven by the camera operator. As he rolls it the actor performs and all the movements get collected under geometrical detections and the motion gets captured. The whole system gets captured by the camera and the motions are all get transformed. Even the props also get captured in the camera and runs through transformation. All the actor’s performances are well recorded and generated for better perception and depiction through the camera. This is the means through which the computer generated characters, along with the present images and sets get the equivalent perspective like those that are generated through video images from the camera. In a computer the data get process and all the display of every particular movement the actor gets synchronized. This is done as per the requirement of the camera position and the script that maintains the plot.
CG. Masi, consider that the whole proceeding of motion tracking or as can be said the motion capture activity gets started as a photogrametric analysis tool. It is basically a leading methodology related to biomechanics research that was under maximum hype during the phase of 1970s and 1980s. Lots of researches were done in this perspective and the periphery was much extended and expanded into the domains related to training, education, sports and in recent years to computer animation for the entertainment business. The basic marketing is done through cinema and video games as they are getting more popularized with the addition of technology maturity. As stated by Fan, J, EA El-Kwae, M-S Hacid, and F Liang (2002, p393) the process to capture the motion begins with the performer who wears markers near each of his joint. This is done in order to identify the motion and to have a clear perception made by the positions or angles between the predestined markers of his joints. These markers are not mere simple markers. They are either of inertial, acoustic, LED, magnetic or something that is called reflective markers. These are used alone or for instance sometimes the combinations of any of these are enough for the purpose of tracking the geometrical dimensional display of the human body in action. This is done optimally at the least in two times and the accountability gets counted by the rate of the desired motion that the human body brings into the camera to a position that is very much identified as a submillimete. The computer software for the purpose of motion capture is very distinctively active for the records of the data collected and for all the geometrical derivation. These are basically distributed and derived from velocities, angles, positions, accelerations and impulses as have been recorded. This is the means that provides an accurate digital representation of all the motions that are been displayed. The same proceeding in terms of biomechanics, sports and training activities acts for the means of deriving real time data. These are all provided through dynamic and very mandatorily stored information. Later proceedings are supported by the diagnose problems or in some instances to the provision of expressing ways to improve performance. This is a act that is in severe position of motion capture technology. The act very dynamically functions for the capture of motions up to 140 miles per hour.
Related Work
There are many works that are being initiated in this particular domain. The closest of all these is the ‘Imitation Learning of Human Movement and Hand Grabbing to Adapt to Environment Changes’ by Stephan Al-Zubi (Universität Kiel, D). Under this research he proposed a model for the purpose of learning the articulated motion of human arm and all those aspects that are related to the act of hand grabbing. The goal of this paper is to generate plausible trajectories of joints of the human body that mimic acts for the human movement through the specific depiction of deformation information. Trajectories are mapped to a constraint space and that initiates the act of configurating the human body and all the task-specific constraints like those of avoiding an obstacle, and more dynamically the acts related to the picking up and putting down objects. This model is meant for the principal component analysis (PCA) and for dynamic cell structure (DCS) network.
‘Contours and Optical Flow: Cues for Capturing Human Motion in Videos’ by Thomas Brox (Universität Bonn, D) deals with the capturing human motion through the collected data from video. This research concentrates on tracking features that are natural to the videos. There must be a regular ignorance to local patches. These also comprises of contours and the optical flow that allow tracking and the pose of objects and humans. This research works towards the supplementing of the optic flow that gets derived for the purpose of large motions. This brings in the complementary feature in general and robust system for human tracking.
Further discussions were initiated in ‘The role of Manifold learning in Human Motion Analysis’ by Ahmed Elgammal (Rutgers Univ. – Piscataway, USA). According to him, ‘human body is an articulated object with high degrees of freedom.’ This is the core idea that he elaborates in his paper and distinctively analyses it on the basis of high dimensionality of the configuration space. He focuses over the issue that many human motion activities gets structured and functions on low dimensional manifolds. All these intrinsic body keep mark of the configuration are subject to manifolds that are low in dimensionality. Ahmed Elgammal discusses about the resulting appearance that again maintain the static quality of manifolds. This is here that the derivative conclusion gets focussed through the challenges that partake in the capturing motions of the model. He no doubt gets the support of all the aspects like those of shape and appearance. There are the provisions that support the person in the act of performing the motion. This is also related to the acts that have got variation in the view point, or in the case related to illumination. This paper bears the whole systematic learning proceedings for the purpose of decomposable generative models. These are again concerned with the subject and to the explicitly decomposing act that the intrinsic body configuration supplies as in the function of time from all kids of sides that are related to the human body under the act. The relations are established in terms of appearance or as can be said the shape of the performer in terms of the time-invariant parameters.
Jürgen Gall (MPI für Informatik – Saarbrücken, D) came up with an exclusive paper entitled as ‘Learning a Dynamic Independent Pose Distribution within a Bayesian Framework’. In his elaborative terms, the act of capturing motion can be elaborated in two steps in a Bayesian framework. He makes a vivid and elaborative study of the applicable poses that are usually captured in the current frame. This again moves out to the periphery of the next frame. The prediction as has been made by Jürgen Gall is that this is a phenomenon that can be well narrated through a dynamic model. This then gets updated through the next frame. The special thing that he interpreted here is related to the human motions that get simplified under the models of the dynamics. This is further is done on the basis for the selection of complexity of the dynamics. The process further proceeds with the determining factors that get integrated prior to the acceptance of knowledge. These are like those of anatomical constraints, self-intersection and above all an exclusive speculation over unrealistic joint. In this paper Jürgen Gall was also very elaborative in making pose distinction orders.
The researches are getting multifaceted and in this perspective the ‘MoCap for Interaction Environments’ by Daniel Grest (Universität Kiel, D) is a perfect elaborative experience. This is a paper related to the facts that gets more idealistically expressed in terms of accuracy marker. It is through this accuracy marker that the motion capture systems gets less comparable state to the marker based systems. However in this paper the most important aspect gets elaborated in case of segmentation step. It is through this segmentation step that strong restrictions to the capture environment gets predominant peripheral domain. These can be well exemplified through homogeneous clothing or the presented background. The constant lighting too is an integral part of this determination. In case of interaction environments the role played by the background is non-static. It further gets cluttered and lighting changes rapidly. Motion capture in these kinds of conditioning environment needs to be supported by robust acts that get the real-time, the accuracy of the image and the footage captured motion. Daniel Grest also discussed exclusively about Stereo algorithms. This is the means through which it can provide robust data with respect to lighting and background. These are represented as they are available in real-time.
Conclusions
Thus to conclude, the new technological participation in the field of human motion capture is getting various dynamic exposure. There is hardly a section that needs well reformation, though the researches are still going on to make new application more effective and string. The whole phenomenon is very interesting and though the depictions a re more human –like then human they are already very popular and the video game parlous and cinema halls are highly benefitted by them. The whole entertainment world appreciates its participation and there is no point in making it something away from the mob. There are still some hurdles persists but with the exceptional finishing ouches the applications are more accomplished and appreciated. Some of the very important aspects that are being looked forward through this paper are need to be further explored. The related research works are the basis that needs to be counted for further innovative exploration of this topic. The dominance as has been derived here are very limited and the further study is very highly recommended. There are innumerable topics to be explored and all these can be systematically positioned through the further readings as has been recommended here. There is thus an open-ended provision led in this paper fro the exploration of this biggest unit and technique of capturing human motion.
References
Aggarwal, JK and Q Cai. (1999) “Human Motion Analysis: A Review.” Computer Vision and Image Understanding 73, no. 3.
Fan, J, EA El-Kwae, M-S Hacid, and F Liang (2002, p393) “Novel tracking-based moving object extraction algorithm.” J Electron Imaging 11.
Green, RD, L Guan, and JA Burne, (2000) “Video analysis of gait for diagnosing movement disorders.” J Electron Imaging 9.
Masi, CG. “Vision improves bat performance.” Vision Systems Design. 2008. Web.
Motion-Based Geometry Compensation for DWT Compression of 3D mesh Sequences Boulfani-Cuisinaud, Y.; Antonini, M. Image Processing, 2007. IEEE International Conference. Digital Object Identifier 10.1109.
Müller, M., Röder, T., Clausen, M. Ecient Content-Based Retrieval of Motion Capture Data. ACM Transactions on Graphics 24(3), (Proceedings of ACM SIGGRAPH 2005).
Shopov, A. et al. (2000) “Improvements in image analysis and fluorescence microscopy to discriminate and enumerate bacteria and viruses in aquatic samples, or cells, and to analyze sprays and fragmenting debris.” Aquatic Microbial Ecology 22.
Sparks, C. et al. (2005) “Comparison and Validation of Smooth Particle Hydrodynamics (SPH) and Coupled Euler Lagrange (CEL) Techniques for Modeling Hydrodynamic Ram.” 46th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, Austin, Texas.