Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 15, 1
eCollection

2DKD: A Toolkit for Content-Based Local Image Search

Affiliations

2DKD: A Toolkit for Content-Based Local Image Search

Julian S DeVille et al. Source Code Biol Med.

Abstract

Background: Direct comparison of 2D images is computationally inefficient due to the need for translation, rotation, and scaling of the images to evaluate their similarity. In many biological applications, such as digital pathology and cryo-EM, often identifying specific local regions of images is of particular interest. Therefore, finding invariant descriptors that can efficiently retrieve local image patches or subimages becomes necessary.

Results: We present a software package called Two-Dimensional Krawtchouk Descriptors that allows to perform local subimage search in 2D images. The new toolkit uses only a small number of invariant descriptors per image for efficient local image retrieval. This enables querying an image and comparing similar patterns locally across a potentially large database. We show that these descriptors appear to be useful for searching local patterns or small particles in images and demonstrate some test cases that can be helpful for both assembly software developers and their users.

Conclusions: Local image comparison and subimage search can prove cumbersome in both computational complexity and runtime, due to factors such as the rotation, scaling, and translation of the object in question. By using the 2DKD toolkit, relatively few descriptors are developed to describe a given image, and this can be achieved with minimal memory usage.

Keywords: Cryo-electron microscopy; Digital pathology; Krawtchouk polynomials; Local descriptors; Local image retrieval; Shape matching; Subimage search.

Conflict of interest statement

Competing interestsThe authors declare no competing interest.

Figures

Fig. 1
Fig. 1
Flow chart of 2DKD. The script names are shown in black boxes
Fig. 2
Fig. 2
Nine small gray-scale clip art images used as subimages to generate the first image dataset. Image credit: Microsoft Office Online – clip art gallery
Fig. 3
Fig. 3
An example 600×600 image from the dataset containing gray-scale clip art subimages
Fig. 4
Fig. 4
Example queries and corresponding retrievals from the dataset with 30% noise. For each query subimage, top 5 matches from the dataset are shown
Fig. 5
Fig. 5
a A section of a projection image of GroEL protein complexes in vitreous ice captured using Cryo-EM. b Averaged top view of GroEL. c Averaged side view of GroEL. d An end-on view of the 3D atomic structure of GroEL complex. Image credits – a Vossman, https://commons.wikimedia.org/wiki/File:Cryoem\_groel.jpg, b, c: Electron Microscopy Data Bank (EMD-8750), d: Protein Data Bank (PDB ID: 5W0S)
Fig. 6
Fig. 6
An example query of the top view of GroEL and top 15 retrieval results using 2DKD. The pixel size for the local subimages is 40×40. The (x,y) centers of the query and retrieval results in the 1024×1024 global image are provided under each subimage

Similar articles

See all similar articles

References

    1. Sit A, Kihara D. Comparison of image patches using local moment invariants. IEEE Trans Image Process. 2014;23(5):2369–79. doi: 10.1109/TIP.2014.2315923. - DOI - PubMed
    1. Sit A, Shin W-H, Kihara D. Three-dimensional Krawtchouk descriptors for protein local surface shape comparison. Pattern Recog. 2019;93:534–45. doi: 10.1016/j.patcog.2019.05.019. - DOI
    1. Sael L, Li B, La D, Fang Y, Ramani K, Rustamov R, Kihara D. Fast protein tertiary structure retrieval based on global surface shape similarity. Proteins Struct Funct Bioinforma. 2008;72(4):1259–73. doi: 10.1002/prot.22030. - DOI - PubMed
    1. Han X, Sit A, Christoffer C, Chen S, Kihara D. A global map of the protein shape universe. PLoS Comput Biol. 2019;15(4):e1006969. doi: 10.1371/journal.pcbi.1006969. - DOI - PMC - PubMed
    1. Mehta N, Raja’S A, Chaudhary V. Content based sub-image retrieval system for high resolution pathology images using salient interest points. In: 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE: 2009. p. 3719–22. 10.1109/iembs.2009.5334811. - PubMed

LinkOut - more resources

Feedback