Talk:Outline of object recognition

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Requested move[edit]

Single object recognitionObject recognitionThe term 'object recognition' is used much more often in computer vision literature and it usually implies 'single object recognition'. Currently the article object recognition redirect to computer vision, which is not good. — Andreas Kaufmann (talk)

I agree on moving this article to object recognition. Tpl (talk) 11:03, 30 March 2008 (UTC)


The article is very misleading. Object recognition is a huge area of research within image processing and computer vision. To talk about one person I've never heard of (no offensive David) kind of gives the wrong impression. The entry should be about object recognition in general rather than talk about one person and his patented method. Doc phil (talk) 12:44, 25 April 2008 (UTC)

I completely agree. David Lowe's SIFT method has become quite popular, but there are many object recognition approaches that do not use it, and object recognition as a field existed for decades before his paper. (talk) 05:43, 7 May 2008 (UTC)

I agree too. I am currently working on object recognition, and the following methods should be mentioned: geometric hashing, local geometric hashing, vocabulary tree... not to mention that there are many other descriptors that are better than SIFT. -- (talk) 02:51, 12 May 2008 (UTC)

I'm a computer vision researcher, and I think this article would be better off deleted. David Lowe is an important and accomplished researcher, but this article reads like an advertisement or biography of his object recognition approaches. If this article wants to objectively mention some of the state of the art methods in object recognition it should look at the PASCAL VOC challenge or the top performers on the MSRC object recognition database. (talk) 17:37, 22 May 2008 (UTC)

The first few sentences could be retained, but the rest should be moved to a page specifically about SIFT. At present this is completely misleading. I suggest referring to some basic courses on object recognition by researchers like Antonio Torralba or Kristen Grauman for accurate information. A better page could probably be synthesized from those fairly easily. (talk) 04:15, 3 March 2009 (UTC)

Yeah, it really seems like an ethusiastic Lowe fan or something has been editing this... -- NIC1138 (talk) 21:43, 13 March 2009 (UTC)

I agree, it describes very little about other methods of object recognition, and people are probably worse off for reading this, re-title the article david lowes object recognition method or something Bogan229 (talk) 02:09, 24 March 2009 (UTC)

New structure of this article[edit]

I think that it was a good suggestion to move the specific material on the SIFT algorithm to the specific article on that topic. To form the basis for a more general article on object recognition, I have added a set of sections on general topics with references to relevant authors. The technical material remains to be filled in. Tpl (talk) 15:58, 12 October 2009 (UTC)

The article is looking a lot better, but as-is needs a lot of work still. We can't possibly hope to list every possible technique here - nor should we, that's what Categories are for. Perhaps instead just a general overview of how object detection algorithms commonly work with eamples of commonly used agorithms. My view is that OR is usually a 2-step process: process an image, then do the recognition. If no one objects, I'll reoganise the middle sections to reflect this to give a general overview and include some of the more common aproaches. --mjog (talk) 02:39, 8 August 2010 (UTC)


This article, object detection and object-class detection seem to me to overlap. Should they all be consolidated here? I don't know the subject well enough to do a merge, so I am just tagging the pages and asking that knowledgable editors have a look. Thanks. EdChem (talk) 14:32, 11 August 2013 (UTC)

Support, except that this article should be Object recognition, which currently redirects here. Lfstevens (talk) 22:16, 9 October 2013 (UTC)
No please don't. Recognition is simply to identify something among similar things of the same family (object class), while detection is to find something among different families (object classes). Please refer to the articles of face recognition and face detection to see the differences, in which the research on faces is a part of the research on objects generally. object detection and object-class detection articles can be merged into one article called Object Detection. (talk) 00:15, 25 December 2013 (UTC)

Quick explanation of Wikipedia outlines[edit]

"Outline" is short for "hierarchical outline". There are two types of outlines: sentence outlines (like those you made in school to plan a paper), and topic outlines (like the topical synopses that professors hand out at the beginning of a college course). Outlines on Wikipedia are primarily topic outlines that serve 2 main purposes: they provide taxonomical classification of subjects showing what topics belong to a subject and how they are related to each other (via their placement in the tree structure), and as subject-based tables of contents linked to topics in the encyclopedia. The hierarchy is maintained through the use of heading levels and indented bullets. See Wikipedia:Outlines for a more in-depth explanation. The Transhumanist 00:08, 9 August 2015 (UTC)