Local quantized extrema patterns for content-based natural and texture image retrieval

Abstract

This paper proposes a novel feature descriptor, named local quantized extrema patterns (LQEP) for content based image retrieval. The standard local quantized patterns (LQP) collect the directional relationship between the center pixel and its surrounding neighbors and the directional local extrema patterns (DLEP) collect the directional information based on local extrema in 0°, 45°, 90°, and 135° directions for a given center pixel in an image. In this paper, the concepts of LQP and DLEP are integrated to propose the LQEP for image retrieval application. First, the directional quantized information is collected from the given image. Then, the directional extrema is collected from the quantized information. Finally, the RGB color histogram is integrated with the LQEP for a feature vector generation. The performance of the proposed method is tested by conducting three experiments on Coel-1K, Corel-5K and MIT VisTex databases for natural and texture image retrieval. The performance of the proposed method is evaluated in terms of precision, recall, average retrieval precision and average retrieval rate on benchmark databases. The results after investigation show a considerable improvements in terms of their evaluation measures as compared to the existing methods on respective databases.

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26 DOI 10.1186/s136730150044z

*Correspondence: [email protected]

1 Department of Electronics and Communication Engineering, JNTU Hyderabad, Hyderabad, Andhra Pradesh, IndiaFull list of author information is available at the end of the article

RESEARCH

Open Access

http://crossmark.crossref.org/dialog/?doi=10.1186/s13673-015-0044-z&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s13673-015-0044-z&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s13673-015-0044-z&domain=pdf

Web End = Local quantized extrema patterns forcontentbased natural andtexture image retrieval

http://crossmark.crossref.org/dialog/?doi=10.1186/s13673-015-0044-z&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s13673-015-0044-z&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s13673-015-0044-z&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s13673-015-0044-z&domain=pdf

Web End = http://crossmark.crossref.org/dialog/?doi=10.1186/s13673-015-0044-z&domain=pdf

Web End = L. Koteswara Rao1* and D. Venkata Rao2

Abstract

This paper proposes a novel feature descriptor, named local quantized extrema patterns (LQEP) for content based image retrieval. The standard local quantized patterns (LQP) collect the directional relationship between the center pixel and its surrounding neighbors and the directional local extrema patterns (DLEP) collect the directional information based on local extrema in 0, 45, 90, and 135 directions for a given center pixel in an image. In this paper, the concepts of LQP and DLEP are integrated to propose the LQEP for image retrieval application. First, the directional quantized information is collected from the given image. Then, the directional extrema is collected from the quantized information. Finally, the RGB color histogram is integrated with the LQEP for a feature vector generation. The performance of the proposed method is testedby conducting three experiments on Coel-1K, Corel-5K and MIT VisTex databases for natural and texture image retrieval. The performance of the proposed method is evaluated in terms of precision, recall, average retrieval precision and average retrieval rate on benchmark databases. The results after investigation show a considerable improvements in terms of their evaluation measures as compared to the existing methods on respective databases.

Keywords: Local quantized patterns (LQP), Ditectional local extrema patterns (DLEP), Texture, Image retrieval, Pattern recognition, Databases

Background

Data mining is an active area of research for mining or retrieving data/information from a large database or library. Image retrieval is the part of data mining in which the visual information (images) from the large size database or library is retrieved. Earlier, text-based retrieval was used for retrieving the information. In this process, the images are annotated with text and then text-based database management systems were used to perform image retrieval. Many advances, such as data modelling, multidimensional indexing, and query evaluation, have been made along this research direction. However, there exist two major difficulties, especially when the size of image collections is too large (tens or hundreds of thousands). One is the vast amount of labor required to annotate the images manually. The other difficulty, which is more essential, araise due to the rich content in the images and the subjectivity of human perception. That is, for

2015 Koteswara Rao and Venkata Rao. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

the same image content dierent people may perceive it dierently. To address these issues, content-based image retrieval (CBIR) came into existence. CBIR utilizes the visual content such as color, texture, shape, etc., for indexing and retrieving images from the database. The comprehensive and extensive literature on feature extraction for CBIR is available in [15].

Color does not only add beauty to images/video but provides more information about the scene. This information is used as the feature for retrieving the images. Various color-based image search schemes have been proposed, some of these are discussed in this section. Swain and Ballard [6] introduced the concepts of color histogram feature and the histogram intersection distance metric to measure the distance between the histograms of the images. The global color histogram is extensively utilized for the purpose of retrieval which gives probability of occurrences of each unique color in the image on a global platform. It is a fast method and has translation, rotation and scale invariant property, but it suers from lack of spatial information which yields false retrieval results. Stricker etal. [7] proposed two new color indexing schemes. In the rst approach they have used the cumulative color histogram. In the second method instead of storing the complete color distributions, rst three moments of each color space of image are used. Idris and Panchanathan [8] have used the vector quantization technique to compress the image and from codeword of each image, they obtained the histogram of codewords which was used as the feature vector. In the same manner, Lu etal. [9] proposed a feature for color image retrieval by employing the combination of discrete cosine transform (DCT) and vector quantization technique. A well known image compression method i.e. block truncation coding (BTC) method is proposed in [10] for extracting two features i.e. block color co-occurrence matrix (BCCM) and block pattern histogram (BPH). Image descriptors are also generated with the help of vector quantization technique. Global histogram also suers from spatial information. In order to overcome this problem Hauang etal. [11] proposed a color correlogram method which includes the local spatial color distribution of color information for image retrieval. Pass and Zabih [12] proposed the color coherence vectors (CCV) where a histogram based approach incorporates some spatial information. Rao etal. [13] proposed the modication in color histogram to achieve the spatial information and for this purpose they proposed three spatial color histograms : annular, angular and hybrid color histograms. Chang etal. [14] proposed a method which takes care of change in color due to change of illumination, the orientation of the surface, and the viewing geometry of the camera with less feature vector length as compared to color correlogram.

Texture is another important feature for CBIR. Smith etal. used the mean and variance of the wavelet coefficients as texture features for CBIR [15]. Moghaddam etal. proposed the Gabor wavelet correlogram (GWC) for CBIR [16]. Ahmadian etal. used the wavelet transform for texture classication [17]. Subrahmanyam etal. have proposed the correlogram algorithm for image retrieval using wavelets and rotated wavelets (WC+RWC)

[18]. Ojala etal. proposed the local binary pattern (LBP) features for texture description [19] and these LBPs are converted to rotational invariant LBP for texture classication [20]. Pietikainen etal. proposed the rotational invariant texture classication using feature distributions [21]. Ahonen etal. [22] and Zhao etal. [23] used the LBP operator facial expression analysis and recognition. Heikkila etal. proposed the background modeling and detection by using LBP [24]. Huang et al. proposed the extended LBP

Page 2 of 24

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

for shape localization [25]. Heikkila etal. used the LBP for interest region description [26]. Li et al. used the combination of Gabor lter and LBP for texture segmentation [27]. Zhang etal. proposed the local derivative pattern for face recognition [28]. They have considered LBP as a nondirectional rst order local pattern, which are the binary results of the rst-order derivative in images. The block-based texture feature which use the LBP texture feature as the source of image description is proposed in [29] for CBIR. The center-symmetric local binary pattern (CS-LBP) which is a modied version of the well-known LBP feature is combined with scale invariant feature transform (SIFT) in [30] to describe the regions of interest. Yao etal. [31] have proposed two types of local edge patterns (LEP) histograms, one is LEPSEG for image segmentation, and the other is LEPINV for image retrieval. The LEPSEG is sensitive to variations in rotation and scale, on the contrary, the LEPINV is resistant to variations in rotation and scale. The local ternary pattern (LTP) [32] has been introduced for face recognition under dierent lighting conditions. Subrahmanyam etal. have proposed various pattern based features such as local maximum edge patterns (LMEBP) [33], local tetra patterns (LTrP) [34] and directional local extrema patterns (DLEP) [35] for natural/texture image retrieval and directional binary wavelet patterns (DBWP) [36], local mesh patterns (LMeP) [37] and local ternary co-occurance patterns (LTCoP) [38] for biomedical image retrieval. Reddy etal. [39] have extended the DLEP features by adding the magnitude information of the local gray values of an image. Hussain and trigges [40] have proposed the local quantized patterns (LQP) for visual recognition.

Recently, the integration of color and texture features are proposed for image retrieval. Jhanwar etal. [41] have proposed the motif co-occurrence matrix (MCM) for content-based image retrieval. They also proposed the color MCM which is calculated by applying MCM on individual red (R), green (G), and blue (B) color planes. Lin et al. [42] combined the color feature, k-mean color histogram (CHKM) and texture features, motif cooccurrence matrix (MCM) and dierence between the pixels of a scan pattern (DBPSP). Vadivel etal. [43] proposed the integrated color and intensity co-occurrence matrix (ICICM) for image retrieval application. First they analyzed the properties of the HSV color space and then suitable weight functions have been suggested for estimating the relative contribution of color and gray levels of an image pixel. Vipparthi etal. [44] have proposed the local quinary patterns for image retrieval.

The concepts of LQP [40] and DLEP [35] have motivated us to propose the local quantized extrema patterns (LQEP) for image retrieval application. The main contributions of this work are summarized as follows. (a) The proposed method collects the directional quantized extrema information from the query/database image by integrating the concepts of LQP and DLEP. (b) To improve the performance of the CBIR system, the LQEP operator combines with RGB color histogram. (c) The performance of the proposed method is tested on benchmark databases for natural and texture image retrieval.

The paper is summarized as follows: In Background, a brief review of image retrieval and related work is given. Review of existing local patterns, presents a concise review of existing local pattern operators. The proposed system framework and similarity distance measures are illustrated in Local quantized extrema patterns (LQEP). Experimental results and discussions are given in **Experimental Results and Discussion. The conclusions and future scope are given in **Conclusion.

Page 3 of 24

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Review ofexisting local patterns

Local binary patterns (LBP)

The LBP operator was introduced by Ojala etal. [19] for texture classication. Success in terms of speed (no need to tune any parameters) and performance is reported in many research areas such as texture classication [1821], face recognition [22, 23], object tracking [33], image retrieval [3339] and nger print recognition. Given a center pixel in the 33 pattern, LBP value is calculated by comparing its gray scale value with its neighboring pixels based on Eqs.(1) and (2):

where I(gc) denotes the gray value of the center pixel, I(gp) represents the gray value of its neighbors, P stands for the number of neighbors and R, the radius of the neighborhood.

After computing the LBP pattern for each pixel (j, k), the whole image is represented by building a histogram as shown in Eq.(3).

where the size of input image is N1 N2.

Figure1 shows an example of obtaining an LBP from a given 33 pattern. The histograms of these patterns contain the information on the distribution of edges in an image.

Block based local binary patterns (BLK_LBP)

Takala et al. [29] have proposed the block based LBP for CBIR. The block division method is a simple approach that relies on subimages to address the spatial properties of images. It can be used together with any histogram descriptors similar to LBP. The

Page 4 of 24

LBPP,R =

P

[notdef]

p=1

2(p1) f1(I(gp) I(gc))

(1)

f1(x) =

[notdef]

1 x 0 0 else

(2)

HLBP(l) =

N1

[notdef]

j=1

N2

[notdef]

k=1

f2(LBP(j, k), l); l [0, (2P 1)]

(3)

f2(x, y) =

[notdef]

1 x = y 0 else

(4)

Fig. 1 Calculation of LBP

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

method works in the following way: First it divides the model images into square blocks that are arbitrary in size and overlap. Then the method calculates the LBP distributions for each of the blocks and combines the histograms into a single vector of sub-histograms representing the image.

Centersymmetric local binary patterns (CS_LBP)

Instead of comparing each pixel with the center pixel, Heikkila etal. [30] have compared center-symmetric pairs of pixels for CS_LBP as shown in Eq.(5):

After computing the CS_LBP pattern for each pixel (j, k), the whole image is represented by building a histogram as similar to the LBP.

Directional local extrema patterns (DLEP)

Subrahmanyam etal. [35] propsed directional local extrema patterns (DLEP) for CBIR. DLEP describes the spatial structure of the local texture based on the local extrema of center gray pixel gc.

In DLEP, for a given image the local extrema in 0, 45, 90, and 135 directions are obtained by computing local dierence between the center pixel and its neighbors as shown below:

The local extremes are obtained by Eq.(7).

The DLEP is dened (=0, 45, 90, and 135) as follows:

Local quantized patterns (LQP)

Hussain and Trigges [40] have proposed the LQP operator for visual recognition. The LQP collects the directional geometric features in horizontal (H), vertical (V), diagonal (D) and antidiagonal (A) strips of pixels; combinations of these such as horizontal-vertical (HV), diagonal-antidiagonal (DA) and horizontalvertical-diagonal-antidiagonal (HVDA); and traditional circular and disk-shaped regions. Figure2 illustrate the possible direction quantized geometric structures for LQP operator. The more details about LQP are available in [40].

Page 5 of 24

CS_LBPP,R =

P

[notdef]

p=1

2(p1) f1(I(gp) I(gp+(P/2)))

(5)

I(gi) = I(gc) I(gi); i = 1, 2, . . . , 8

(6)

~(gc) = f3(I(gj), I(gj+4)); j = (1 + ~

~45) ~ = 0, 45, 90, 135

(7)

f3(I(gj), I(gj+4)) =

[notdef]

1 I(gj) I(gj+4) 0 0 else

(8)

DLEP(I(gc))

[notdef][notdef]

=

(gc);(g1);(g2); . . .(g8)

(9)

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 6 of 24

Fig. 2 The possible directional LQP structures

Local quantized extrema patterns (LQEP)

The operators DLEP [35] and LQP [40] have motivated us to propose the LQEP for image retrieval. The LQEP integrates the concepts of LQP and DLEP for image retrieval. First, the possible structures are extracted from the give pattern. Then, the extrema operation is performed on the directional geometric structures. Figure3 illustrates the calculation of LQEP for a given 77 pattern. For easy understanding of the readers, in Fig. 3 the 7 7 pattern is indexed with pixel positions. The positions are indexed in a manner to write the four directional extrema operator calculations. In this paper,

HVDA7 geometric structure is used for feature extraction. The brief description about the LQEP feature extraction is given as follows.

Fig. 3 The LQEP calculation for a given 7 7 pattern using HVDA7 geometric structure

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

For a given center pixel (C) in an image I, the HVDA7 geometric structure is collected using Fig.3. Then the four directional extremas (DE) in 0, 45, 90, and 135 directions are obtained as follows.

where,

The LQEP is dened by Eq.(10)(13) as follows.

Eventually, the given image is converted to LQEP map with values ranging from 0 to 4095. After calculation of LQEP, the whole image is represented by building a histogram supported by Eq.(16).

where, LQEP(i,j) represents the LQEP map value ranging 0 to 4095.

Proposed image retrieval system

In this paper, we integrate the concepts of DLEP and LQP for image retrieval. First, the image is loaded and converted into gray scale if it is RGB. Secondly, the four directional HVDA7 structure is collected using the LQP geometric structures. Then, the four directional extremas in 0, 45, 90, and 135 directions are collected. Finally, the LQEP feature is generated by constructing the histograms. Further, to improve the performance of the proposed method, we integrate the LQEP with color RGB histogram for image retrieval.

Figure4 depicts the owchart of the proposed technique and algorithm for the same is given as follows.

Algorithm:Input: Image; Output: Retrieval result

1. Load the image and convert into gray scale (if it is RGB).2. Collect the HVDA7 structure for a given center pixel.3. Calculate the local extrema in 0, 45, 90, and 135 directions.

Page 7 of 24

DE(I(gc))

[notdef][notdef]

0 =

f4(I(g1), I(g4), I(gC)); f4(I(g2), I(g5), I(gC)); f4(I(g3), I(g6), I(gC))

(10)

DE(I(gc))

[notdef][notdef]

45 =

f4(I(g13), I(g16), I(gc)); f4(I(g14), I(g17), I(gc)); f4(I(g15), I(g18), I(gc))

(12)

(11)

DE(I(gc))

[notdef][notdef]

90 =

f4(I(g7), I(g10), I(gC)); f4(I(g8), I(g11), I(gC)); f4(I(g9), I(g12), I(gC))

(13)

DE(I(gc))

[notdef][notdef]

135 =

f4(I(g19), I(g22), I(gC)); f4(I(g20), I(g23), I(gC)); f4(I(g21), I(g24), I(gC))

f4(x, y, c) =

1 if (x > c) or (y > c) 1 if (x < c) or (y < c) 0 else

(14)

LQEP =

[notdef]

DE(I(gc))

0 , DE(I(gc))

45 , DE(I(gc))

90 , DE(I(gc))

135

(15)

HLQEP(l) =

N1

[notdef]

j=1

N2

[notdef]

k=1

f2(LQEP(j, k), l); l [0, 4095];

(16)

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 8 of 24

Fig. 4 Proposed image retrieval system framework

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

4. Compute the 12-bit LQEP with four directional extrema.5. Construct the histogram for 12-bit LQEP.6. Construct the RGB histogram from the RGB image.7. Construct the feature vector by concatenating RGB and LQEP histograms.8. Compare the query image with the image in the database using Eq.(20).9. Retrieve the images based on the best matches

Query matching

Feature vector for query image Q is represented as fQ = (fQ1, fQ2, . . . fQLg ) is obtained after the feature extraction. Similarly each image in the database is represented with feature vector fDBj = (fDBj1, fDBj2, . . . fDBjLg ); j = 1, 2, . . . , |DB|. The goal is to select n best images that resemble the query image. This involves selection of n top matched images by measuring the distance between query image and image in the database DB. In order to match the images we use four dierent similarity distance metrics as follows.

where, Lg represents the feature vector length, I1 is database image and fDBji is ith feature of jth image in the database DB.

Experimental results anddiscussion

The performance of the proposed method is analyzed by conducting two experiments on benchmark databases. Further, it is mentioned that the databases used are Corel-1K [45], Corel-5K [45] and MIT VisTex [46] databases.

In all experiments, each image in the database is used as the query image. For each query, the system collects n database images X=(x1, x2,, xn) with the shortest image matching distance computed using Eq.(20). If the retrieved image xi=1, 2,, n belongs to same category as that of the query image then we say the system has appropriately identied the expected image else the system fails to nd the expected image.

The performance of the proposed method is measured in terms of average precision/ average retrieval precision (ARP), average recall/average retrieval rate (ARR) as shown below:

Page 9 of 24

Manhattan distance measure: D(Q, I1) =

Lg

[notdef]

i=1

fDBji fQ,i

(17)

1/2

Euclidean distance measure: D(Q, I1) =

Lg

i=1

(fDBji fQ,i)2

(18)

fDBji fQ,i

Canberra distance measure: D(Q, I1) =

Lg

[notdef]

i=1

(19)

fDBji

+

f

Q,i

d1 distance measure: D(Q, I1) =

Lg

[notdef]

i=1

fDBji fQ,i1 + fDBji + fQ,i

(20)

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 10 of 24

For the query image Iq, the precision is dened as follows:

Experiment #1

In this experiment, Corel-1K database [44] is used. This database consists of large number of images of various contents ranging from animals to outdoor sports to natural images. These images have been pre-classied into dierent categories each of size 100 by domain professionals. Some researchers think that Corel database meets all the requirements to evaluate an image retrieval system, due its large size and heterogeneous content. For experimentation we selected 1000 images which are collected from 10 different domains with 100 images per domain. The performance of the proposed method is measured in terms of ARP and ARR as shown in Eq.(2124). Figure5 illustrates the sample images of Corel-1K database.

Table1 and Fig.6 illustrate the retrieval results of proposed method and other existing methods in terms of ARP on Corel1K database. Table2 and Fig.7 illustrate the retrieval results of proposed method and other existing methods in terms of ARR on Corel1K database. From Tables1, 2, Figs.6 and 7, it is clear that the proposed method

Precision: P(Iq) = Number of relevant images retrieved

Total Number of images retrieved

(21)

Average Retrieval Precision: ARP = 1

|DB|

|DB|

[notdef]

i=1

P(Ii)

(22)

Recall: R(Iq) = Number of relevant images retrieved

Total Number of relevant images in the database

(23)

Average Retrieval Rate: ARR = 1

|DB|

|DB|

[notdef]

i=1

R(Ii)

(24)

Fig. 5 Sample images of Corel-1K database

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Table 1 The retrieval results ofproposed method andvarious other existing method interms ofARP atn=20 onCorel-1K database

Page 11 of 24

posed

method

Africans53.1568.380.4078.2569.7552.4657.2055.3560.9061.3061.9575.15

Beaches43.855441.2544.2554.2551.3343.6052.0553.9051.2553.957.65

Buildings48.756.255.6559.1063.9555.6663.3562.2563.4057.8561.9574.7

Buses82.888.876.7086.0589.6596.3395.5095.896.5594.4097.494.3

Dinosaurs9599.399.0098.7098.7095.2396.894.598.0097.8598.7598.95

Elephants34.8565.856.2059.0048.8042.0346.0543.3546.1048.905056.55

Flowers88.3589.192.9085.3592.3085.6391.4085.2086.6089.1091.8595.45

Horses59.3580.376.5074.9589.45065.3064.7569.4072.1566.2077.386.65

Mountains30.852.233.7036.5547.3035.9334.5533.5536.1039.4043.7545.9

Food50.473.370.6064.4070.9070.3670.6576.1575.0575.3578.6582

Total58.7272.768.2968.6672.565.0366.3866.7668.8768.1671.5576.73

LBPLTPLDPLTrPReddy etal.LQPPro

CategoryAverage retrieval precision (ARP) (%); (n=20)

Jhanwar etal.Lin etal.CCVadivel etal.Subrahmanyam

etal.

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 12 of 24

Fig. 6 Comparison of proposed method with other existing methods in terms of ARP on Corel-1K database

shows a signicant improvement as compared to other existing methods in terms of precision, ARP, recall and ARR on Corel1K database. Figure8a, b illustrate the analysis of proposed method (LQEP) with various similarity distance measures on Corel-1K database in terms of ARP and ARR respectively. From Fig.8, it is observed that the d1 distance measure outperforms the other distance measures in terms of ARP and ARR on Corel-1K database. Figure9 illustrates the query results of proposed method on Corel 1K database.

Experiment #2

In this experiment Corel-5K database [45] is used for image retrieval. The Corel-5K database consists of 5000 images which are collected from 50 dierent domains have 100 images per domain. The performance of the proposed method is measured in terms of ARP and ARR as shown in Eq.(2124).

Table3 illustrates the retrieval results of proposed method and other existing methods on Corel-5K database in terms of precision and recall. Figure10a, b show the category wise performance of methods in terms of precision and recall on Corel-5K database. The performance of all techniques in terms of ARP and ARR on Corel-5K database can be seen in Fig.10c, d respectively. From Table3, Fig.10, it is clear that the proposed method shows a signicant improvement as compared to other existing methods in

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Table 2 The retrieval results ofproposed method andvarious other existing method interms ofARR atn=20 onCorel-1K database

Page 13 of 24

posed

method

Africans32.2142.146.2948.4143.5838.132.938.138.639.2541.2344.38

Beaches29.0432.125.2925.8535.7735.429.436.238.333.8234.6838.32

Buildings27.736.535.0137.0534.8933.73536.534.931.9633.5544.84

Buses48.6661.760.9766.5263.3970.569.974.273.473.5777.8573

Dinosaurs81.4494.189.5978.1192.7875.187.577.283.790.2892.1692.55

Elephants21.4233.134.1435.6630.3125.427.828.529.530.5330.9132.78

Flowers63.537577.6957.7364.5965.671.362.265.869.3276.579.99

Horses35.8447.636.1341.4766.5542.240.444.343.136.1647.850.39

Mountains21.7527.721.0224.3732.0926.923.624.627.529.3529.5532.32

Food29.024939.2738.2445.1237.240.547.952.245.352.2554.54

Total39.0649.8946.5445.3450.9144.945.846.948.747.9551.6454.31

LBPLTPLDPLTrPReddy etal.LQPPro

Jhanwar etal.Lin etal.CCVadivel etal.Subrahmanyam

etal.

CategoryAverage Retrieval Recall (ARR) (%)

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 14 of 24

Fig. 7 Comparison of proposed method with other existing methods in terms of ARR on Corel-1K database

terms of their evaluation measures on Corel-5K database. The performance of the proposed method is also analyzed with various distance measures on Corel-5K database as shown in Fig.11. From Fig.11, it is observed that the d1 distance measure outperforms the other distance measures in terms of ARP and ARR on Corel-5K database. Figure12 illustrates the query results of the proposed method on Corel-5K database.

Experiment #3

In this experiment, MIT VisTex database is considered which consists of 40 dierent textures [46]. The size of each texture is 512512 which is divided into sixteen 128128 non-overlapping sub-images, thus creating a database of 640 (4016) images. In this experiment, each image in the database is used as the query image. Average retrieval

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 15 of 24

Fig. 8 Comparison of proposed method with various distance measures in terms of ARP and ARR on Corel-1K database

recall or average retrieval rate rate (ARR) given in Eq.(24) is the bench mark for comparing results of this experiment.

Figures13, 14 illustrate the performance of various methods in terms of ARR and ARP on MIT VisTex database. From Figs.13, 14, it is clear that the proposed method shows a signicant improvement as compared to other existing methods in terms of ARR and ARP on MIT VisTex database. Figure15 illustrates the performance of proposed method with dierent similarity distance measures in terms of ARR on MIT VisTex database. From Fig. 15, it is observed that d1 distance measure outperforms the other distance measures in terms of ARR on MIT VisTex database. Figure 16 illustrates the query results of proposed method on MIT VisTex database.

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 16 of 24

Fig. 9 Query results of proposed method on Corel-1K database

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 17 of 24

Table 3 Results ofvarious methods interms ofprecision andrecall onCorel-5K database

Corel5 KPrecision (%)32.941.535.145.743.648.854.450.762.2

Recall (%)14.018.314.820.319.221.124.122.429.6

DatabasePerformanceMethod

CS_LBPLEPSEGLEPINVBLK_LBPLBPDLEPReddy etal.LQPPM

PMproposed method, BLK_LBPblock based LBP [25]

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 18 of 24

Fig. 10 Comparison of proposed method with other existing methods on Corel5K. a Category wise performance in terms of precision, b category wise performance in terms of recall, c total database performance in terms of ARP and d total database performance in terms of ARR

Fig. 11 Comparison of proposed method with various distance measures in terms of ARP on Corel-5K database

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 19 of 24

Fig. 12 Query results of proposed method on Corel-5K database (top left imageis the query image)

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 20 of 24

Fig. 13 Comparison of proposed method with other existing methods in terms of ARR on MIT VisTex database

Conclusions

A new feature descriptor, named local quantized extrema patterns (LQEP) is proposed in this paper for natural and texture image retrieval. The proposed method integrates the concepts of local quantization geometric structures and local extrema for extracting features from a query/database image for retrieval. Further, the performance of the proposed method is improved by combining it with the standard RGB histogram. Performance of the proposed method is tested by conducting three experiments on benchmark image databases and retrieval results show a signicant improvement in terms of their evaluation measures as compared to other existing methods on respective databases.

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 21 of 24

Fig. 14 Comparison of proposed method with other existing methods in terms of ARP on MIT VisTex database

Fig. 15 Comparison of proposed method with various distance measures in terms of ARR on MIT VisTex database

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 22 of 24

Fig. 16 Query results of the proposed method on MIT VisTEx database (top left imageis the query image)

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

Page 23 of 24

Authors contributions

All authors are contributed equally for the current research work. All authors read and approved the nal manuscript.

Author details

1 Department of Electronics and Communication Engineering, JNTU Hyderabad, Hyderabad, Andhra Pradesh, India.

2 Narasaraopet Institute of Technology, Guntur, Andhra Pradesh, India.

Acknowledgements

The authors declare that they have no acknowledgements for the current research work.

Compliance with ethical guidelines

Competing interests

The authors declare that they have no competing interests.

Received: 13 December 2014 Accepted: 11 August 2015

References

1. Aura C, Castro EMMM (2003) Image mining by content. Expert Syst Appl 23:3773832. Subrahmanyam M, Maheshwari RP, Balasubramanian R (2012) Expert system design using wavelet and color vocabulary trees for image retrieval. Expert Syst Appl 39:51045114

3. Vipparthi SK, Nagar SK (2014) Expert image retrieval system using directional local motif XoR patterns. Expert Syst Appl 41(17):80168026

4. Rui Y, Huang TS (1999) Image retrieval: current techniques, promising directions and open issues. J Vis Commun Image Represent 10:3962

5. Smeulders AWM, Worring M, Santini S, Gupta A, Jain R (2000) Content-based image retrieval at the end of the early years. IEEE Trans Pattern Anal Mach Intell 22(12):13491380

6. Swain MJ, Ballard DH (1991) Color Indexing. Int J Comput Vision 7:11327. Stricker M, Orengo M (1995) Similarity of color images. In: Proceedings SPIE Storage and Retrieval for Image and Video Databases III, San Jose, Wayne Niblack, pp 381392

8. Idris F, Panchanathan S (1997) Image and video indexing using vector quantization. Mach Vis Appl 10:43509. Lu ZM, Burkhardt H (2005) Colour image retrieval based on DCT domain vector quantisation index histograms. Electron Lett 41:956957

10. Qiu G (2003) Color image indexing using BTC. IEEE Trans Image Process 12:9310111. Huang J, Kumar S, Mitra M, Zhu W, Zabih R (1997) Image indexing using color correlograms. In: Proceeding Computer Vision and Pattern Recognition, San Jaun, Puerto Rico, pp 762768

12. Pass G, Zabih R (1996) Histogram renement for content-based image retrieval. In: Proceedings IEEE Workshop on Applications of Computer Vision, 1996, pp 96102

13. Rao A, Srihari RK, Zhang Z (1999) Spatial color histograms for content-based image retrieval. In: Proceedings of the Eleventh IEEE International Conference on Tools with Articial Intelligence, Chicago, IL, USA 1999, pp 183187

14. Chang MH, Pyun JY, Ahmad MB, Chun JH, Park JA (2005) Modied color co-occurrence matrix for image retrieval. Lect Notes Comput Sci 3611:4350

15. Smith JR, Chang SF (1996) Automated binary texture feature sets for image retrieval. In: Proceedings IEEE International Conference on Acoustics, Speech and Signal Processing, Columbia Univ., New York, pp 22392242

16. Moghaddam HA, Khajoie TT, Rouhi AH (2003) A new algorithm for image indexing and retrieval using wavelet correlogram. In: International Conference on Image Processing, K.N. Toosi Univ. of Technol., Tehran, Iran, vol. 2, pp 497500

17. Ahmadian A, Mostafa A (2003) An efficient texture classication algorithm using Gabor wavelet. In: 25th Annual international conf. of the IEEE EMBS, Cancun, Mexico, pp 930933

18. Murala S, Maheshwari RP, Balasubramanian R (2011) A correlogram algorithm for image indexing and retrieval using wavelet and rotated wavelet lters. Int J Signal Imaging Syst Eng 4(1):2734

19. Ojala T, Pietikainen M, Harwood D (1996) A comparative sudy of texture measures with classication based on feature distributions. J Pattern Recogn 29(1):5159

20. Ojala T, Pietikainen M, Maenpaa T (2002) Multiresolution gray-scale and rotation invariant texture classication with local binary patterns. IEEE Trans Pattern Anal Mach Intell 24(7):971987

21. Pietikainen M, Ojala T, Scruggs T, Bowyer KW, Jin C, Homan K, Marques J, Jacsik M, Worek W (2000) Overview of the face recognition using feature distributions. J Pattern Recogn 33(1):4352

22. Ahonen T, Hadid A, Pietikainen M (2006) Face description with local binary patterns: applications to face recognition. IEEE Trans Pattern Anal Mach Intell 28(12):20372041

23. Zhao G, Pietikainen M (2007) Dynamic texture recognition using local binary patterns with an application to facial expressions. IEEE Trans Pattern Anal Mach Intell 29(6):915928

24. Heikkila M, Pietikainen M (2006) A texture based method for modeling the background and detecting moving objects. IEEE Trans Pattern Anal Mach Intell 28(4):657662

25. Huang X, Li SZ, Wang Y (2004) Shape localization based on statistical method using extended local binary patterns. In: Proceedings of International Conference on Image and Graphics, pp 184187

26. Heikkila M, Pietikainen M, Schmid C (2009) Description of interest regions with local binary patterns. J Pattern recognition 42:425436

Koteswara Rao and Venkata Rao Hum. Cent. Comput. Inf. Sci. (2015) 5:26

27. Li M, Staunton RC (2008) Optimum Gabor lter design and local binary patterns for texture segmentation. J Pattern Recogn 29:664672

28. Zhang B, Gao Y, Zhao S, Liu J (2010) Local derivative pattern versus local binary pattern: Face recognition with higher-order local pattern descriptor. IEEE Trans Image Proc 19(2):533544

29. Takala Valtteri, Ahonen Timo, Pietikainen Matti (2005) Block-based methods for image retrieval using local binary patterns, SCIA 2005. LNCS 3450:882891

30. Heikkil Marko, Pietikainen M, Schmid C (2009) Description of interest regions with local binary patterns. Pattern Recogn 42:425436

31. Yao Cheng-Hao, Chen Shu-Yuan (2003) Retrieval of translated, rotated and scaled color textures. Pattern Recogn 36:913929

32. Tan X, Triggs B (2010) Enhanced local texture feature sets for face rec-ognition under difficult lighting conditions. IEEE Trans Image Process 19(6):16351650

33. Murala S, Maheshwari RP, Balasubramanian R (2012) Local maximum edge binary patterns: a new descriptor for image retrieval and object tracking. Signal Process 92:14671479

34. Murala S, Maheshwari RP, Balasubramanian R (2012) Local tetra patterns: a new feature descriptor for content based image retrieval. IEEE Trans Image Processing 21(5):28742886

35. Murala S, Maheshwari RP, Balasubramanian R (2012) Directional local extrema patterns: a new descriptor for content based image retrieval. Int J Multimedia Inf Retrieval 1(3):191203

36. Murala S, Maheshwari RP, Balasubramanian R (2012) Directional binary wavelet patterns for biomedical image indexing and retrieval. J Med Syst 36(5):28652879

37. Murala S, Jonathan Wu QM (2013) Local mesh patterns versus local Binary patterns: biomedical image indexing and retrieval. IEEE J Biomed Health Inform 18(3):929938

38. Murala S, Jonathan Wu QM (2013) Local ternary co-occurrence patterns: a new feature descriptor for MRI and CT image retrieval. Neurocomputing 119(7):399412

39. Reddy PVB, Reddy ARM (2014) Content based image indexing and retrieval using directional local extrema and magnitude patterns. Int J Electron Commun (AE) 68(7):637643

40. ul Hussain S, Triggs B (2012) Visual Recognition Using Local Quantized Patterns. ECCV 2012, Part II, LNCS 7573, Italy, pp 716729

41. Jhanwar N, Chaudhuri S, Seetharaman G, Zavidovique B (2004) Content-based image retrieval using motif cooccurrence matrix. Image Vision Comput 22:12111220

42. Lin CH, Chen RT, Chan YKA (2009) Smart content-based image retrieval system based on color and texture feature. Image Vision Comput 27:658665

43. Vadivel A (2007) Sural Shamik, Majumdar AK. An integrated color and intensity co-occurrence matrix. Pattern Recognit Lett 28:974983

44. Vipparthi SK, Nagar SK (2014) Color directional local quinary patterns for content based indexing and retrieval. Human Centric Comp Inform Sci 4:6. doi:http://dx.doi.org/10.1186/s13673-014-0006-x

Web End =10.1186/s13673-014-0006-x

45. Corel 1000 and Corel 10000 image database. [Online]. Available: http://wang.ist.psu.edu/docs/related.shtml

Web End =http://wang.ist.psu.edu/docs/related.shtml 46. MIT Vision and Modeling Group, Vision Texture. [Online]. Available: http://vismod.media.mit.edu/pub/

Web End =http://vismod.media.mit.edu/pub/

Page 24 of 24