- 1、原创力文档(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。。
- 2、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载。
- 3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
- 4、该文档为VIP文档,如果想要下载,成为VIP会员后,下载免费。
- 5、成为VIP后,下载本文档将扣除1次下载权益。下载后,不支持退款、换文档。如有疑问请联系我们。
- 6、成为VIP后,您将拥有八大权益,权益包括:VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
- 7、VIP文档为合作方或网友上传,每下载1次, 网站将根据用户上传文档的质量评分、类型等,对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档
查看更多
大数据和视觉识别:潜力和挑战-ZhangLei2012-09-25-10_25_14
Big Data for Visual Recognition:
Potentials and Challenges
Lei Zhang / Senior Researcher
Joint work with colleagues: Xin-Jing Wang, Changhu Wang, Zhiwei Li, Richard Cai,
Kuiyuan Yang, and interns: Xiao Zhang, Yang Cao, Zhenbang Sun, Qiang Hao
Microsoft Research Asia
September 21, 2012
A Dream To Understand Images
鈥 Image 飪 Word?
Vision is the process of discovering from images what is present in the world,
and where it is.
- David Marr
Visual Recognition: Model-Based Approach
for General Category
Model-Based
Approaches
飪 Complex
飪 Train for each category
飪 Win by algorithms
Challenges in Model-based Approaches
鈥 Great intra-class variance
Visual Recognition:
Specific Instance vs. General Category
CN Tower
Eiffel Tower
Data Driven Model-Based
Approaches Approaches
Leifeng Tower
飪 Simple 飪 Complex
飪 Generalizable to any objects 飪 Train for each category
飪 Win by scalability 飪 Win by algorithms
Space needle
Utilizing the Power of Data
鈥 Efficient image search and matching
痰 More reliable similarity measure
鈥 For two images, which part is similar to which part?
痰 Efficient nearest neighbor search
Big Data and Recognition: State of the Art
鈥 Image Annotation and Visual Recognition
痰 Search-based image annotation
痰 Instance-based visual recognition
鈥 Visual Pattern Mining from Massive Data
鈥 Building Image Knowledge Base by Matching
Images wit
文档评论(0)