4 ms·

Algorithm to enhance low-res muscle mesh with high-res muscle mesh in real-time

24 points by flacle 11y ago

mentos 11y ago
What are the applications for this?
- flacle 11y ago
  For this specific project/paper it's any application that would include real-time computer animation with an underlying musculoskeletal framework. Biomechanics scientists, clinicians can use it to improve the "visual" realism of their bio-simulation tools, or video game developers and artists can use it to increase the realism of game characters. The benefit here is that you can let the model takes care of the biomechanics and it's output is used to deform the muscle and tendon geometry in real-time. The number of vertices (or resolution) of each muscle can also be adjusted to keep the minimum required frame rate, in case you want dynamic skinning and everything else that needs to be added into a complex scene.
austinjp 11y ago
Very interesting. What are the ramifications for improving the understanding of biomechanics? Does this allow for better investigation into the effects of deformations such as compression, or muscles that are bent over pivot points? Are cross-sectional areas easily available, and can these be used to determine if muscle power is affected in certain situations eg the deformations mentioned?
- flacle 11y ago
  Any improvement in biomechanics would have a lot of positive consequences as the field has a lot of subfields. One example is soft body dynamics. Research in this area is very active as computational power increases more accurate and complex finite element simulations of soft tissues are possible. While the classical model used in the paper, from Delp et al., already contains pivot points, it lacks other important info. For instance, the tendon slack length parameter currently includes both the length of the free tendon and the length of the aponeurotic tendon. In reality most musculotendons consist of at least one tendon on each side. AFAIK, no functional representation is available that divides the tendon length into separate lengths for the origin and insertion of each unit, and also into its two constituents, free and aponeurotic. And as has been shown, deformation effects for muscles and tendon are indeed different as they they have different material properties. Regarding the cross-sectional areas, these are available, just not within the classical models. If there is a need to incorporate cross-sectional area into a model than the main issue is measure. Most models in biomechanics are created by taking external measurements of subjects in vivo, but I guess it won't take long before models go beyond piecewise line segments. Or soft body simulation get cheap enough to reach the same level of accuracy present in these trial-and-tested models.

24 points by flacle 11y ago

mentos 11y ago

What are the applications for this?

flacle 11y ago
For this specific project/paper it's any application that would include real-time computer animation with an underlying musculoskeletal framework. Biomechanics scientists, clinicians can use it to improve the "visual" realism of their bio-simulation tools, or video game developers and artists can use it to increase the realism of game characters. The benefit here is that you can let the model takes care of the biomechanics and it's output is used to deform the muscle and tendon geometry in real-time. The number of vertices (or resolution) of each muscle can also be adjusted to keep the minimum required frame rate, in case you want dynamic skinning and everything else that needs to be added into a complex scene.

austinjp 11y ago

Very interesting. What are the ramifications for improving the understanding of biomechanics? Does this allow for better investigation into the effects of deformations such as compression, or muscles that are bent over pivot points? Are cross-sectional areas easily available, and can these be used to determine if muscle power is affected in certain situations eg the deformations mentioned?

flacle 11y ago
Any improvement in biomechanics would have a lot of positive consequences as the field has a lot of subfields. One example is soft body dynamics. Research in this area is very active as computational power increases more accurate and complex finite element simulations of soft tissues are possible. While the classical model used in the paper, from Delp et al., already contains pivot points, it lacks other important info. For instance, the tendon slack length parameter currently includes both the length of the free tendon and the length of the aponeurotic tendon. In reality most musculotendons consist of at least one tendon on each side. AFAIK, no functional representation is available that divides the tendon length into separate lengths for the origin and insertion of each unit, and also into its two constituents, free and aponeurotic. And as has been shown, deformation effects for muscles and tendon are indeed different as they they have different material properties. Regarding the cross-sectional areas, these are available, just not within the classical models. If there is a need to incorporate cross-sectional area into a model than the main issue is measure. Most models in biomechanics are created by taking external measurements of subjects in vivo, but I guess it won't take long before models go beyond piecewise line segments. Or soft body simulation get cheap enough to reach the same level of accuracy present in these trial-and-tested models.