Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Ett tillvägagångssätt för extrahering och igenkänning av objekt
2014 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

There are limited methods for finding information in a video, such as manually added metadata, using face recognition or text recognition. This master thesis will propose a new method for extraction and recognition of objects from color images. Through this, the frames in video material can be described in a new dimension, the content of the video. This makes the video content searchable and enables description by more than just the provided metadata.Even though there exists a lot of research in the field of object recognition there is no clear solution to the problem of extracting objects that can be recognized. To solve this problem, different image processing techniques are used for analyzing images, or frames in a video. The approach presented is built up in three steps.The first step is based upon edge detection as a means of segmenting the image into different parts. A new method is proposed as the second step, where contours are identified and extracted from the edges detected by the first step. The last step uses signal analysis and Fourier descriptors for indexing and comparing different shapes. The shapes are normalized and filtered to reduce the required signature size and improve the invariant properties for comparison to other shapes.The chosen edge detection gives clear edges and a straight forward approach into the proposed method of finding contours. The proposed method finds many shapes with good precision. With signal analysis applied, the shapes can be indexed and compared to other shapes. It is shown that on a set of test data, this thesis work will recognize shapes with a good precision and is comparable to other recognition solutions. Also presented are suggestions and a discussion about how to improve the detection of shapes from more properties than this thesis has covered.

Place, publisher, year, edition, pages
2014. , 45 p.
Keyword [en]
Technology
Keyword [sv]
Teknik, object extraction, object recognition, fourier transform
Identifiers
URN: urn:nbn:se:ltu:diva-58943Local ID: f7dfc662-ac39-4605-b15b-c6bce346b33fOAI: oai:DiVA.org:ltu-58943DiVA: diva2:1032331
Subject / course
Student thesis, at least 30 credits
Educational program
Computer Science and Engineering, master's level
Supervisors
Note
Validerat; 20140916 (global_studentproject_submitter)Available from: 2016-10-04 Created: 2016-10-04Bibliographically approved

Open Access in DiVA

fulltext(3431 kB)89 downloads
File information
File name FULLTEXT02.pdfFile size 3431 kBChecksum SHA-512
f8727052bcb33f7ab12864be4d5604467dba8d454359b4996f27daf5ac35b1ae65451f05bf12fa91ad993c2276eaeec0745368c31c18ab824c2e9ca72dc0ff53
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Kornhammar, Tim

Search outside of DiVA

GoogleGoogle Scholar
Total: 89 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 25 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf