II. RELATED WORK
While to the best of our knowledge finding and indexing of
explicit spatial relations between pairs of images for image
retrieval has not been a topic in research yet, there are several
efforts closely related to and fundamental for this work. The bag
of visual words approach, where similar local features are cluster
to build a visual word vocabulary, takes image overlap implicitly
into account as pictures with overlapping content are assumed
similar [3][4]. However, finding and indexing an explicit pair
wise spatial relation, like indexing that image A shows a scene
left of image B, has not been discussed in this context. In image
registration on the other hand focuses on finding image overlap.
Brown [7] and Zitova and Flusser [12] give a great overview of
current methods. Inferring a spatial relationship and indexing this
relationship, however, has not been discussed.
For both, the bag of visual words approach for image retrieval
and image registration, local features build the basis for further
research. Lowe [10] and Bay et al. [5] explain how local features,
which are invariant to translation, rotation and scale, work and
how important these features are to identify objects within an
image or create stitched images. A comparison of different global
and local feature algorithms has been done by Deselaers et al. [9].
Bauer et al. [1] compared the SIFT (short for Scale Invariant
Feature Transform) and the SURF (short for Speeded-Up Robust
Features) algorithm. They identified the SURF algorithm as the
fastest and most reliable algorithm in practical tests. Applications
of SURF include for instance the work of Bay et al. [6] who
created an application employing the SURF algorithm to identify
exhibits in a museum. Brown and Lowe [8] used their local
feature detection algorithm to stitch images automatically.
Snavely et al. used local features to reconstruct a three
dimensional scene [11].
Integrated Module: (How to implement HCI and OOP or coding part together)
Implementation plan:
1 Create project and project details
2 Proving our idea with the user requirements
3 Design graphical user interface
4 chose developing tools
5 start writing the program
6 Trial the program
7 Using the program by user
8 Present the program
Technological plan (requirement) : HTML5