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。

Ask HN: How does Touch ID in the new iMacs work?

2 点作者 trwhite大约 4 年前
Asking out of mere curiosity. Apple&#x27;s new iMacs give you the ability to switch user by scanning your fingerprint.<p>I assume that there&#x27;s some code that converts your fingerprint into some sort of unique identifier but I&#x27;m wondering how it does this? How is the fingerprint scan data stored?<p>All answers&#x2F;suggestions welcome.

3 条评论

moooo99大约 4 年前
I don&#x27;t think the new TouchID works any different than the versions before. According to Apple, the fingerprints are exclusively stored on the device, so, using a locally stored fingerprint for log in a local user should be fairly straight forward.<p>When I got my S10 and everyone talked about the new fingerprint sensor technology I got curious and researched myself, I found a Computerphile [1] video that does a great job at explaining how fingerprint recognition works.<p>[1] <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=xD88Qs_DZp4" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=xD88Qs_DZp4</a>
评论 #26891822 未加载
fundamental大约 4 年前
In case you&#x27;re more curious about fingerprint recognition itself the basic classical approach is roughly:<p>First you acquire a fingerprint image from the sensor. A fingerprint is composed of small ridges and valleys. Every ridge will make physical contact with the sensing device in some way, so that allows for the system to obtain a black and white image showing depth. From there the standard features that are detected is when the ridges either end or split into two ridges (think of them as corners in other image recognition tasks).<p>In order to obtain the location and direction of the feature points (referred to as minutia) the system algorithms usually segment the image into fingerprint&#x2F;noise, performs some signal enhancements, changes it into a skeleton (i.e. each ridge is represented by a 1px thick line), and then gets the location+angle of minutia. Matching can have a number of additional features collected, but assuming the process is as described you can find a similarity score by finding the best rotational+translational fitting from two inputs and seeing how many pairs of minutia line up. From there it&#x27;s just a matter of putting a threshold on the similarity score to decide if it&#x27;s a match.
评论 #26891840 未加载
runjake大约 4 年前
We don&#x27;t know yet. Soonish, the Apple Security Guide[1] should be updated with the correct answer.<p>My best guess is that the keyboard itself is running it&#x27;s own Secure Enclave and is running a modern incantation of BridgeOS and it uses a pairing model[2] to the host Mac.<p>1. <a href="https:&#x2F;&#x2F;support.apple.com&#x2F;guide&#x2F;security&#x2F;welcome&#x2F;web" rel="nofollow">https:&#x2F;&#x2F;support.apple.com&#x2F;guide&#x2F;security&#x2F;welcome&#x2F;web</a><p>2. <a href="https:&#x2F;&#x2F;support.apple.com&#x2F;guide&#x2F;security&#x2F;pairing-model-security-secadb5b6434&#x2F;1&#x2F;web&#x2F;1" rel="nofollow">https:&#x2F;&#x2F;support.apple.com&#x2F;guide&#x2F;security&#x2F;pairing-model-secur...</a>