<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/1312.6082/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:16:22 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/1312.6082/feed.xml" rel="self" type="application/rss+xml" /><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by cbhl</title><link>https://news.ycombinator.com/item?id=13767463</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:13767463</guid><pubDate>Wed, 01 Mar 2017 19:39:56 GMT</pubDate><description>&lt;p&gt;The state of the art has improved so much that for the older tasks, computers are arguably better than humans at them: https://arxiv.org/abs/1312.6082 So humans have to clear higher and higher hurdles to prove themselves human while the computers use the answers from the captchas to get closer to humans&amp;#x27; capabilities.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by kastnerkyle</title><link>https://news.ycombinator.com/item?id=11912139</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:11912139</guid><pubDate>Wed, 15 Jun 2016 21:05:40 GMT</pubDate><description>&lt;p&gt;Sure, there are a lot of potential things to do - either by enumerating towards the full &amp;quot;brute force&amp;quot; loss partway as I think you are hinting toward, or treating as a subsequences, or something else [0, 1, 2]. The point is, IMO RNNs do need some kind of structured loss (more than per step likelihood) to be competitive with HMM approaches using Viterbi decoding, in addition to some sequence smarts (such as beam search) in decoding. At least in NMT, enumerating the possible decisions as 3n + 1 is almost impossible since the softmax size is generally the memory bottleneck in training - and a…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by cbhl</title><link>https://news.ycombinator.com/item?id=11067060</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:11067060</guid><pubDate>Tue, 09 Feb 2016 17:37:19 GMT</pubDate><description>&lt;p&gt;Distorted text isn&amp;#x27;t even used in reCAPTCHA all that much anymore -- Google used a Deep Convolutional Neural Network and got over 90% accuracy in reading street numbers in Street View, and 99.8% accuracy in solving distorted text reCAPTCHA (http://arxiv.org/abs/1312.6082). Most users will just check a &amp;quot;I&amp;#x27;m not a robot&amp;quot; box now; and if you do get a test, it will likely be a computer vision / image labeling problem: https://googleonlinesecurity.blogspot.com/2014/12/are-you-ro...&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by daviddumenil</title><link>https://news.ycombinator.com/item?id=8047400</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:8047400</guid><pubDate>Thu, 17 Jul 2014 13:37:18 GMT</pubDate><description>&lt;p&gt;The original paper being discussed: http://arxiv.org/abs/1312.6082&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by yaroslavvb</title><link>https://news.ycombinator.com/item?id=7951198</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:7951198</guid><pubDate>Thu, 26 Jun 2014 19:29:09 GMT</pubDate><description>&lt;p&gt;We found that neural networks can solve CAPTCHAS much better than humans, 99.8% on the &amp;quot;hard&amp;quot; ReCAPTCHA instances: http://arxiv.org/pdf/1312.6082.pdf This is why visual recognition is just one of the signals you need to use to tell humans and computers apart http://googleonlinesecurity.blogspot.com/2014/04/street-view...&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by spcoll</title><link>https://news.ycombinator.com/item?id=7602090</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:7602090</guid><pubDate>Thu, 17 Apr 2014 05:12:04 GMT</pubDate><description>&lt;p&gt;It&amp;#x27;s a new success for Deep Learning. It seems to be actually 99.8% accuracy according to their paper: http://arxiv.org/abs/1312.6082 That&amp;#x27;s one order of magnitude higher.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by dwiel</title><link>https://news.ycombinator.com/item?id=7601018</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:7601018</guid><pubDate>Wed, 16 Apr 2014 22:52:09 GMT</pubDate><description>&lt;p&gt;http://arxiv.org/abs/1312.6082&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by pavelrub</title><link>https://news.ycombinator.com/item?id=7600668</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:7600668</guid><pubDate>Wed, 16 Apr 2014 21:30:49 GMT</pubDate><description>&lt;p&gt;This is essentially the technology that was discussed here 3 months ago [1], and it links to the exact same article on arxiv, titled: &amp;quot;Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks&amp;quot;. [2] They new addition to the article is that now they have tested the same type of NN on reCAPTCHA, and (perhaps unsurprisingly) it works. [1] - https://news.ycombinator.com/item?id=7015602 [2] - http://arxiv.org/abs/1312.6082v4.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item><item><title>Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks — mentioned by magicalist</title><link>https://news.ycombinator.com/item?id=7015961</link><guid isPermaLink="false">urn:hackerpapers:mention:1312.6082:7015961</guid><pubDate>Tue, 07 Jan 2014 06:24:01 GMT</pubDate><description>&lt;p&gt;I think the actual paper (linked at the end) is much more informative and interesting than this coverage: http://arxiv.org/abs/1312.6082 (I wouldn&amp;#x27;t call it blogspam, because it looks like they interviewed the researchers, but the summary leaves something to be desired)&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1312.6082/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item></channel></rss>
