Computational Bodybuilding: Anatomically-Based Modeling of Human Bodies | Hacker News Reader

Computational Bodybuilding: Anatomically-Based Modeling of Human Bodies

dl.acm.org

32 pointsjcr9 comments

This seems pretty crude. Particularly in that it ignores visceral fat entirely. Seems like a huge sacrifice in the accuracy of the output. The only acceptable application I could think of for a model with this level of detail would be for EA's "The Sims" game.

They did acknowledge everything you mentioned in part (5). Not sure why you feel the need make this comment.

Meta discussion is one of the core strengths of HN comments. If the GP felt that this was the most important takeaway from the article, then it's good that s/he didn't force others to read 9 dense pages of prose in order to gain exposure to the idea. And really, the title mentioning bodybuilding implies that this is a practical model. The GP is pointing out that it probably isn't.

You are not seeing this situation clearly.

The research presented is useful for far more than models for the "EA Sims game" as was disparagingly claimed. The title is also fitting since the methods presented are applicable to quickly building practical models of bodybuilding. Though the modeling of visceral/organ fat is specifically excluded for further research, the fat modeling methods presented are both effective and useful for model generation. Their use of a quasi-static solver ("Projective Dynamics") on both their muscle and fat tissue growth modeling to get "near-interactive" run times is particularly impressive. The single paragraph "conclusion" section of the paper shows the paper holds interesting, useful, and innovative research:

>"To our knowledge, our work is the first to simulate physics-based growth processes of human tissues in computer graphics. We believe that our system will be instrumental in reducing the often prohibitive costs of human body modeling and will find applications even beyond the traditional realms of computer graphics, such as film, games, and visual effects."

Of course, a lot of people won't have the time to read a long academic paper, so comment quantity and quality often suffers on submissions of academic works, but there's a problem here that you seem to be missing.

The comment by 'sigmar' is actually just another example of the long standing "mean and/or dumb" comment problem on HN [1]. The comment is both factually wrong and unnecessarily derogatory. That comment should have been at least down-voted, and possibly even flagged, rather than defended as some imaginary "core strength" of HN.

The phrase "mean and/or dumb" is from PG, but it's a old and common problem affecting most discussion forums. Manipulative people regularly make failed attempts to look smart in public discussions by being excessively negative, and sadly, it tends to work in so much as lots of other people are fooled into up-voting them. Anyone can be a totally uninformed and acrimonious critic, and unfortunately, most human beings are complete suckers for controversy, so the "mean and/or dumb but massively up-voted" comment problem tends to be self-reinforcing. The only good news is, both the bad commenters and the bad up-voters are quietly handled in code [2].

When you see the "mean and/or dumb" comments, or even the excessively negative comments, you should down-vote them since they are against the site guidelines [3, 4]. When a comment is particularly egregious, click the "X hours ago" link on the comment, and then the 'flag' link (if it's available). Like the down-vote privilege, the flag privilege also has a karma threshold to prevent abuse.

When you realize how the excessively negative comments are really just trying to seek attention, you understand they are just another form of trolling. In this case, 'sigmar' is doing simple drive-by-trolling; say something as offensive as possible to get attention, and refuse to back it up with references when questioned about it. Yep, you got trolled. But I hope you'll try to look for this sort of negative nonsense in the future, and down-vote/flag it appropriately.

[1] https://news.ycombinator.com/item?id=2403696

[2] https://news.ycombinator.com/newsfaq.html

[3] https://news.ycombinator.com/newsguidelines.html

[4] https://news.ycombinator.com/item?id=9317916

There is trolling happening here, however, it's not coming from sigmar who offered a terse criticism and moved on. In that light, this will be my final post on this thread.

I put in the effort to disagree with you politely and provide a detailed explanation with references. Though it seems you've written off my response as trolling, I hope you remember and reconsider it. You'll be better off in the long run if you do.

>In this case, 'sigmar' is doing simple drive-by-trolling; say something as offensive as possible to get attention, and refuse to back it up with references when questioned about it.

How is the game reference "offensive"? Their own article offered up "games" as an application of the research (in the section you quoted).

Furthermore, no one asked me a single question about my comment.

You must have a ton of time on your hands to go around trolling like this.

Wow, that's some great research there. We are actually building a scale for consumers that will allow for body scans in 3D. We hope then to also later integrate morphing algorithms so that you can see how your body would transform if you were to gain extra muscles and fat. Potentially big data could also help.

As for the visceral fat, it's indeed there, but an MRI could probably help off-set such, if you wanted to have a high level of accuracy.

For most I believe outside fat will be most important.

Great title.

Reply on news.ycombinator.com