• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Concloud
 

Concloud

on

  • 626 views

Concloud project general presentation

Concloud project general presentation

Statistics

Views

Total Views
626
Views on SlideShare
622
Embed Views
4

Actions

Likes
0
Downloads
3
Comments
0

2 Embeds 4

http://www.slideshare.net 3
https://duckduckgo.com 1

Accessibility

Categories

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Concloud Concloud Presentation Transcript

    • ATMOS based content retrival library. The Concloud project
    • Preface Content retrival and media indexing systems developing has become a bit of cultwords in the last years. Today we can observe a number of attempts to create some retrival systems, presented as web services, database buildups etc. That can't make us to stand that content retrival problem remains largely unsolved. We hope, we have been heading for right trend to solve some problems while on developing Concloud poject.
    • Let's combine the trends Cloud computing is an extensive conception in data manipulation for today. It is rather intenseand convinient solution both for storaging andcomputing. Drawing in of cloud computing by some big internet commerce market participant, such as Amazon has shown already opulance of such technology. EMC ATMOS provides all the preferences of cloud computing. Here we have: convinient api, simple system of metadata based retrival and classification, high level of reliability and safety .
    • Let's combine the trends What would it be alike, if we base our content retrival system on ATMOS? Will we gain some perfection? Firstly, we will chalange many problems with simple data manipulation due to convinient metadata managment. Its simple, it have good performance, its safety. Secondly, we will add more universality in our content retrival. For example: henceforth we are able to build retrival either on machine content indexing or on manual tagging each madia object.
    • What would it look like?
      • *It's universal. We have wrapped our system in convinient system of libraries, that can be attached to every web services or application
      • *It's simple to use. Really simpe, try and make sure!
      • *It is on leading edge of algorithms.
    • Where can I use it right now Ok, if you are developing something for domains that sounds like:
      • * Art collections
      • * Photograph archives
      • * Retail catalogs
      • * Medical diagnosis
      • * Crime prevention
      • * Intellectual property
      • * Architectural and engineering design
      • .
    • Where can I use it right now
      • * Geographical information and remote sensing systems
      • * Social networking
      • * Media services
      • * Something so cool, that you definitly can speak about only to investors.
    • Where can I use it right now While the advantages of usage of content retrival in projects like remote sensing systems, photograph archieves and social networking is rather evidental, the usage of it in web commerce and Intellectual property can involve some new ideas, that you can observe in other parts of our presentation. See part applications.
    • How does it look like?
      • All weakly perceptable image noises are ignored.
      • The same shape and edges points that images are suspiciously similar.
      For today we present you a library that provides interface capable to upload picture to ATMOS and retrive images by passing sample picture passed by user. Our system then returns all pictures visually and even cognotively similar to samle. Here we have some requirements our algorithm satisfy:
    • How does it look like?
      • If shapes are not geometrically same, but we know that they are close, algorithm takes it into account. For example «A» and «А» are really close.
      • We may consider knowledges about picture, given by our users.
      • The same pictures are... same.
      • An amount of ajustments can be provided to user to optimize retrival. If user simply wants
    • What we are expecting to develop?
      • Video. Retrival of content from video.
      • Faces. More flexible work with faces.
      • Computation of content difference.