TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

FastMRI initiative releases neuroimaging data set

125 点作者 moneil971超过 5 年前

12 条评论

axegon_超过 5 年前
&gt; Apply for Access<p>&gt; The application process includes acceptance of the Data Sharing Agreement (found below) and submission of an online application form. The application must include the investigator’s institutional affiliation and the proposed uses of the data. NYU fastMRI data may be used for internal research or educational purposes only as described in the data use agreement and may not be redistributed in any way without prior permission.<p>&gt; Read and agree to the data use agreement below to apply for access.<p>Seriously?!?!?!?!?! You call that open source?!??!?! OK, let&#x27;s leave the semantics aside for a moment. Yes, a dataset like that would be very interesting and I would happily play around with it for a few weeks and see what I can come up with. If it&#x27;s something worth exploring further, I&#x27;ll happily document it and open source it. But that isn&#x27;t something I can really estimate without being to explore the data and see first hand what it is. For that reason alone, myself (and dozens off the top of my head) will roll their eyes and pretend like it doesn&#x27;t exist.<p>False claims, over-hyping with no real understanding and bureaucratic crap like this is what is slowing everyone and everything down, the sooner people understand it, the better.
评论 #21776298 未加载
评论 #21775939 未加载
dontreact超过 5 年前
I will reiterate my comment on these types of projects. The regulatory pathway established by the fda for these types of products is woefully inadequate and they are very very hard to properly validate.<p>I think any application of deep convolutional neural networks should be alongside a radiologist. If we speed up scans and make up for it with convnets it is very hard (practically speaking: impossible) to properly validate that they will not hallucinate away rare abnormalities. It will also be impossible for radiologists your spot errors like this in the wild because of the reduction in quality of the scan.<p>What happens when the scanners change their behavior in some subtle way that is unaccounted for by FastMRI? It could start erasing a ton of subtle abnormalities and this would not be possible to check for since the original scan will be lower quality.
评论 #21775848 未加载
评论 #21776566 未加载
评论 #21776928 未加载
评论 #21779275 未加载
评论 #21775692 未加载
ChrisFoster超过 5 年前
Current MR physicist &#x2F; data scientist here. There seems to be a lot of misapprehension in this thread.<p>First, this work is about taking data in the sensor domain (&quot;k-space&quot;) and reconstructing it into an image. Doing this with partial k-space data and hand-coded heuristics is a <i>completely standard</i> part of the MRI research agenda and has been for quite some time. See, for example, <a href="http:&#x2F;&#x2F;mriquestions.com&#x2F;k-space-trajectories.html" rel="nofollow">http:&#x2F;&#x2F;mriquestions.com&#x2F;k-space-trajectories.html</a>. Further, several of these techniques have already made it into routine clinical work, and this acquisition-side stuff generally happens before the radiologist even sees the image (reliable acquisition is in the interaction of radiographer with the scanner manufacturer&#x27;s software).<p>There&#x27;s also various claims here that seem to imply learned reconstruction inherently implies the risk of hallucinations without recourse. Naturally, one should be careful about this, but it&#x27;s just a matter of careful cross validation: hold out examples of abnormal anatomy for the test set. There&#x27;s other ways to attack this problem too: training can be done partly or mostly on synthetic data because we have reasonably good forward models of the physics. In this case, one could choose a wide variety of arbitrary synthetic anatomies during training, to further ally the fear of always hallucinating the &quot;typical human brain&quot; from any scan.<p>Slow acquisition and image artifacts in MRI are a fact of life for people in the field and I believe there&#x27;s huge scope for improvement if we had more intelligent reconstruction and acquisition. Ideally the reconstruction would feed dynamically back into the acquisition to gather more context as necessary; the MR machine is, after all, one giant programmable physics experiment. This is already done in a limited way, but in what I&#x27;ve seen it relies on a lot of hand-coded heuristics. And guess what&#x27;s the logical step after hand-coded heuristics? Yes, learned models where you objectively optimize for a final result, rather than hand-coding based on a few examples.<p>Final note - publicly releasing human data is a massive effort in data cleaning and careful anonymization. Not to mention that the acquisition of each sample is extraordinarily expensive. So bravo to these guys for going to the effort.
lvs超过 5 年前
This is a typical misapplication of machine learning. It&#x27;s important to realize what information can possibly be learned by a trained network. In this case, the only thing that can be learned is which components of Fourier space can be ignored for a given imaging problem. Such a question is far more rationally addressed by a deterministic algorithm, if the goal is to speed up acquisition for any specific anatomy. But to generally assume that an imaging protocol can ignore parts of frequency&#x2F;phase space without generating artifacts is not only wrong, but very dangerous for patients. I can easily play with Fourier space and generate the appearance of pathological conditions that don&#x27;t exist -- or disappear ones that do! Not good!
评论 #21779225 未加载
y-c-o-m-b超过 5 年前
As someone with an undiagnosed neurological illness, it would be fun if I could run my MRI backups through it.
评论 #21774930 未加载
评论 #21787945 未加载
评论 #21777232 未加载
andbberger超过 5 年前
Fantastic!! Accelerating MRI with ML is an idea I&#x27;ve had in my little idea book for years and I&#x27;m delighted to see it getting some mainstream attention!<p>It&#x27;s a serious technical challenge but the benefits could be enormous.<p>IIRC the vast majority of the cost of an MRI is the amortized cost of the imager, so faster scans should hopefully directly reduce the cost to patients, perhaps to the point that regular full-body MRI scans for preventative healthcare could be feasible.
评论 #21775386 未加载
评论 #21775748 未加载
markmiro超过 5 年前
Would they be filling in missing data with AI? Would they be doing something similar to DeepFovea? If so, I would be concerned about accuracy.<p><a href="https:&#x2F;&#x2F;ai.facebook.com&#x2F;blog&#x2F;deepfovea-using-deep-learning-for-foveated-reconstruction-in-ar-vr&#x2F;" rel="nofollow">https:&#x2F;&#x2F;ai.facebook.com&#x2F;blog&#x2F;deepfovea-using-deep-learning-f...</a>
Copenjin超过 5 年前
Is this accessible only to people affiliated with some research institution as it seems[1]?<p>[1] <a href="https:&#x2F;&#x2F;fastmri.med.nyu.edu&#x2F;" rel="nofollow">https:&#x2F;&#x2F;fastmri.med.nyu.edu&#x2F;</a>
评论 #21774091 未加载
pflats超过 5 年前
Huh, I was part of a study during my brain MRIs at NYU Lagone. I wonder if my brain is in there.
d-d超过 5 年前
... are these people aware these images of their bodies are publicly available?
caycep超过 5 年前
Was this at NEURIPS?
MaupitiBlue超过 5 年前
Get ready to learn to code diagnostic radiologists.