英语翻译摘要|个人照片正以一种加速的比率被以数字形式记录下来,而我们用于搜索、浏览和分享这些照片的计算工具正努力的跟上步

英语翻译
摘要|个人照片正以一种加速的比率被以数字形式记录下来,而我们用于搜索、浏览和分享这些照片的计算工具正努力的跟上步伐.一种有前途的方法是自动人脸识别,这将使照片根据它们所包含的个人的特征被分类.然而,在网络规模实现精确的识别,需要在亿万个个体之间进行区分,这似乎是一项艰巨的任务.本文认为,社会网络环境可能是大规模的人脸识别成功的关键.通过在线社交网络站点,许多个人照片进行了网络共享,对于这些共享的图像,我们可以调整这样的社会网络的资源和结构,以提高人脸识别率.利用来自于一个流行的在线社会网络成员中的志愿者的真实照片集,我们模拟仿真了资源的可用性,以改善人脸识别,并讨论了应用这些资源的技术.
关键词|人脸识别;图形模式;社会网络环境;结构预测
一个索引照片的有效方法——特别是个人照片——是通过它们所包含个人的身份,而且在理论上,这可以使用自动人脸识别大规模地执行.然而,从人的面部图像识别个体是一个很难的问题,尤其是当这些图像就像是在图1和图2中所示：
收集于野外,并且姿态,照明和表情都不受控制的变化着.这种困难在大型在线照片集剧增,其中数亿人可能会同时出现；对于任何一个特定的个人的外观变化,个人之间在外观的差异就相对的变得非常小.
此外,即使是为处理数据所做的准备（例如,手动标记图像）,对于在自动识别系统中的人们来说,也会成为沉重的负担.
本文认为,对于在网络上的大型照片集来说,在线社交网络将会成为人脸识别成功的关键.
这一论点是基于两点意见.
首先,网上社区诱导会员有着广泛的动机来对人脸图像手动附加身份标签.这种实践结果,即用户自愿地“标记”他们自己和他们的朋友的照片,能产生超大量标记好的面部图像,从而减少或消除了传统的录入负担.
第二个发现是,一个在线社区的社会网络图,它通常表现为机器可读的形式,提供了强大的上下文信息,既提高了性能,又提高了计算效率.
通过提取包含在在线社会网络人脸书中的照片,我们评估了标记的人脸数据的可用性,并且建立在我们早期的研究基础之上[2],来说明怎么调节社会网络环境,以改善识别.虽然这些结果是初步的,它们也表明“全社会性的”人脸识别是一个值得研究重视的问题.
在联机社会网络里的一个重要的信息源——特别是人脸书——是已被手动标记,或“标上”身份标签的大量的人脸图像集.
标记的受欢迎有点令人惊讶,因为通过关联图像说明、注释或关键字来标记图像是一个乏味的过程——如此的乏味,以至于很少有人真正花时间来标记在他们的个人图书馆里的图像[23].
这种缺乏个人标记的情况仍然存在,尽管事实上,有效率的标签会使个人照片集更协调,但这种情况几乎存在了近10年[24] - [26],而标签可以明显地改善个人图像的组织和检索[27].
有趣的是,当图像在网上进行共享以后,情况似乎有所改变.网上社区诱导社会性的动机去标记,而且,人脸书和其他网上社区中标记的密度体现了这点,这种动机相当强劲.
最近的研究也开始探讨这一现象[28] - [30],它们暗示着社会性的标记诱因可以是相当多样化的.在人脸书中,一副图像中的标签通常对应于个人的身份,而这些标记是用来保证图像会被这个人的朋友所看到.
当艾弗里在一本人脸书的照片上标记了本,本收到了一封邮件信息,链接到这张图片,于是艾弗里和本的朋友们都可能会在这个网站上的新闻供稿数据流里发现以上提到的这个标签.
这样,艾弗里成功地与本分享了照片,并且（或许）本的形象也会在他们的朋友圈中得到加强.
以上,译成英文.

ABSTRACT | Personal photographs are being captured in
digital form at an accelerating rate, and our computational
tools for searching, browsing, and sharing these photos are
struggling to keep pace. One promising approach is automatic
face recognition, which would allow photos to be organized by
the identities of the individuals they contain. However,
achieving accurate recognition at the scale of the Web requires
discriminating among hundreds of millions of individuals and
would seem to be a daunting task. This paper argues that social
network context may be the key for large-scale face recognition
to succeed. Many personal photographs are shared on the
Web through online social network sites, and we can leverage
the resources and structure of such social networks to improve
face recognition rates on the images shared. Drawing upon real
photo collections from volunteers who are members of a
popular online social network, we asses the availability of
resources to improve face recognition and discuss techniques
for applying these resources.
One useful way to index photographs V especially personal
Photographs V is through the identities of the individuals
they contain, and, in theory, this can be executed at
scale using automatic face recognition. However, recognizing
individuals from facial images is a hard problem,
particularly when the images are like those in Figs. 1 and 2:
collected Bin the wild[ with uncontrolled variations in
pose, lighting, and expression. This difficulty is exacerbated
in large online photo collections in which hundreds of
millions of individuals might appear; the difference in appearance
between individuals becomes very small relative
to the appearance variation of any particular individual.
Furthermore, even the preparation of training data (by
manually labeling images, for example) to enroll people in
an automatic recognition system becomes burdensome.
This paper argues that online social networks can
provide the keys to successful face recognition in large
photo collections on the Web. This argument is based on
two observations. First, online communities induce social
incentives for members to manually attach identity labels
to facial images. The resulting practice of users voluntarily
B tagging[ themselves and their friends in photos can
produce extraordinary quantities of labeled facial images,
which reduces or eliminates the traditional enrollment
burden. The second observation is that the social network
graph of an online community, which is often available in
machine-readable form, provides powerful contextual
information that improves both performance and computational
efficiency.
By drawing on photos embedded in the online social
network Face book, we assess the availability of labeled face
data, and we build on our earlier study [2] to show how
social network context can be leveraged to improve
recognition. While these results are preliminary, they
suggest that B socially aware[ face recognition is a problem
that deserves research attention.
One important source of information in online social
Networks V and Face book in particular V is the vast quantity
of facial images that have been manually labeled, or
B tagged, [ by identity. The popularity of tagging is somewhat
surprising, because tagging images by associating
captions, annotations, or keywords is a tedious process V so
tedious that very few people actually take the time to tag the
images in their personal libraries [23]. This lack of personal
tagging persists despite the fact that efficient tagging
interfaces for personal photo collections have existed for
almost a decade [24]–[26] and tags can significantly
improve personal image organization and retrieval [27].
Interestingly, things seem to change when images are
shared online. Online communities induce social incentives
to tag, and, as evidenced by the density of tags in Face book
and other online communities, these incentives can be
quite strong. Recent studies are beginning to explore this
phenomenon [28]–[30], and they suggest that the social
incentives for tagging can be quite diverse. On Face book,
tags typically correspond to the identities of individuals in
an image, and these tags are used to ensure that an image
will be seen by one’s friends. When Avery tags Ben in a
Face book photo, Ben receives an e-mail message with a
link to the image, and both Avery’s friends and Ben’s
friends might find the tag mentioned in streaming B news
feeds[ on the site. In this way, Avery successfully shares
the photo with Ben, and (perhaps) Ben’s stature is
enhanced among their combined group of friends.
Whatever the reasons for social tagging, the practice is a
boon for recognition systems. At the time of this writing,
Face book has a rapidly growing population of more than
400 million users, and it hosts over 20 billion images, with
more than 2.5 billion new photos being added every month
[31], [32]. Many of these images have been manually tagged
with individuals’ identities, and, in this way, the members of
this online community have inadvertently created an
astoundingly large database of annotated facial images
embedded in a social network structure that can be accessed
(at least partially) in machine readable form.
我们老师正好讲这篇文章,中英文都有,如果需要,全篇可以发给你.