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。

Using Deep Learning to Help Pathologists Find Tumors

89 点作者 rusht将近 7 年前

9 条评论

whafro将近 7 年前
I work in this field (not directly on the ML, for the company born out of the winner of Camelyon16) and the last two years of progress has been amazing to watch. Tumor detection has become incredibly accurate, across basically every tissue&#x2F;tumor type, and we&#x27;re now making real progress on the next major goal: determining the best therapy for a given patient.<p>It&#x27;s a bit of a dirty secret in this space that pathologists have a pretty high error rate on a lot of these tasks — it&#x27;s just tough work for human eyes to do literally hundreds of times every day. Applying computer vision techniques can not only improve accuracy and reproducibility over human assessment, but you can do types of analysis in seconds-to-minutes that would literally take years for a human. We&#x27;re just scratching the surface.<p>There are lots of ML challenges here, but just as many general tech&#x2F;engineering&#x2F;design challenges. So if you&#x27;re interested in working on bringing work like this to the masses, we&#x27;d love to talk at PathAI.
评论 #17364893 未加载
评论 #17363893 未加载
评论 #17389797 未加载
mkstowegnv将近 7 年前
I went to a talk by someone who had switched fields to one that involved analyzing portions of cells. After showing a series of slides with diverse, confusing blobs and lines, he said &quot;when I first started this work I would look at a section and not see anything at all. But I have improved to the point that now I can look at a section and see anything I want to&quot;.
toolslive将近 7 年前
I did a project like this early 2000s, and it&#x27;s amazing how far you get by just combining frequency filtering and knn-clustering. Nothing fancy required. really.
评论 #17364496 未加载
评论 #17364071 未加载
Gatsky将近 7 年前
At the moment, a big limitation of this approach is the input data. Images of tumours are generally very thin sections of a complex 3D tissue that is processed in a way that introduces artefacts and then stained with 2 colours.<p>To truly leverage the power of machine learning, an end to end solution where the tissue is processed in a more data rich manner would be better (eg spatially aware single cell assays, non destructive thick slice imaging). This would feasibly replace the current system entirely, as it truly would do something no human could do, not just do it more accurately.
phonebucket将近 7 年前
While open sourcing the model is nice, it would be better still to open source the data set for the wider community to make more meaningful contributions.<p>Their GitHub repo states the following: &quot;You need to apply for data access, and once it&#x27;s approved, you can download from either Google Drive, or Baidu Pan.&quot;
评论 #17367176 未加载
louden将近 7 年前
It would be nice to see the sensitivity and specificity of the technique and for humans. False positives and false negatives are not equal in medicine, so we should report in such a way that people can evaluate them.<p>In this type of cancer, a lower specificity is an acceptable trade off for a very high sensitivity.
评论 #17367113 未加载
sooheon将近 7 年前
How is the &quot;grid of patches&quot; different from one more level of convolution?
评论 #17363291 未加载
评论 #17367229 未加载
评论 #17363155 未加载
评论 #17363022 未加载
godzillabrennus将近 7 年前
Glad to see they open sourced their work.
leozou将近 7 年前
great work