Advertisement

Salient object detection using the phase information and object model

  • Hooman AfshariradEmail author
  • Seyed Alireza Seyedin
Open Access
Article
  • 87 Downloads

Abstract

One of the most important features of saliency detection algorithms is to reduce the size of processing data for algorithms with higher processing size such as object detection algorithms. A main condition for algorithms of saliency detection to be used in detecting the object in the image is their low processing size and broadness of the application extent while having acceptable precision. In this article we introduce a Salient Object Detection method using Task Simulation (SOD-TS). This method has a low processing size and wide functional domain using task simulation (object model). Our proposed method has a wide range of application including ship detection, words and letter detection in texts, etc. Relying on the task simulation (object model), SOD-TS method detects the salient object which is the best response to the current task. It uses the information of the frequency domain.

Keywords

Salient regions detection Fourier transform (FT) phase Frequency domain Object detection Ship detection Optical character recognition (OCR) Task simulation 

1 Introduction

The main reason for using visual attention algorithms is the reduction of the considerable size of data for higher level processes that usually require heavy processes. Approximately 108 bits of information enter into the human visual system. The online processing of such large information is impossible (or at least very costly) without being helped by from visual attention [8]. The use of visual attention leads to process of considerably smaller size of data such that the processed regions would include more significant and exciting information with more information than the adjacent regions. Salient region detection has several applications as explained by Borji, et al. (2013) among which we can refer to object detection, object extraction, image segmentation, tracking, etc.

In terms of the used models, saliency detection methods are classified in two groups: Top-Down (TD) attention and Bottom-Up (BU) attention. In BU (i.e. the sense-driven) methods, lower-level features such as the color, intensity and orientation are used for detecting the salient regions, while in TD (i.e. expectation-driven) methods, cognitive phenomena such as the expectations, knowledge and response to current goals are used as the characteristic of higher level features. Accessibility of lower-level features is one of the advantages of BU methods in which the region in question is indeed a region whose lower-level features are different from other regions. Such methods are specifically suitable for the images with a single salient object; but if we have several salient objects in the image, BU methods detect the most salient one while the people with different interests may perceive different objects as the most salient in the image. For example, in an image with a car and a bicycle. Even if the car is more salient in the image, a bicycler may perceive the region of the bicycle more saliently. TD methods don’t face with such a weakness because their algorithm detect the region in question based on the current task. Indeed, the algorithm detects a region with the best response to the current task. But a main problem of TD methods lays in modelling the cognitive phenomena such as the task. The algorithms using TD methods have a higher processing level because they use neural networks [47] and/or deep learning [29] to model the cognitive phenomena, or they extract the features to find the salient regions such that they fulfill the task simulation process.

One of the broadest applications of visual attention is the object detection. In this process, first the regions with more probability of object presence are specified by regions detection methods; then at the mentioned region the intended object is specified using higher level algorithms of object detection which have higher processing size. The intended regions are detected by several methods each of which has its own strengths and weaknesses. A main problem of such methods is that have either high speed with low precision or high precision with high processing size. The high processing size is an important challenge for such methods because the objective of the methods of salient region detection for the object detection in the image is to reduce the data size for higher level processes of object detection. Since object detection algorithms undertake high size processes, if we want to reduce the processing data size that are heavier than the higher level processes, then the useful feature of such algorithms is neutralized in practice. Thus, the main objective of using such algorithms is to reduce the data size for object detection algorithms while the processing size of salient region detection algorithm must be lower than the object detection algorithm with acceptable precision.

On the other hand, the proposed method of this article (SOD-TS method hereafter) uses the information of frequency domain to simulate the task with very low processing size, and considers the obtained simulated task as one of the algorithm input; then considering this specified task, the algorithm detect the salient object as the best response to the current task. Since our proposed method uses the information of frequency, in following sections first we review two fast and famous methods (SR and PFT methods) that are similar to our proposed and use the information of frequency domain. Then we will introduce SOD-TS method in more details.

2 Frequency domain information

Fourier Transform (FT) phase of the signal includes very important information. Indeed, when the signal has an important part, the values of the signal phase are higher. Aiger, et al. (2010) have used this point to find defects on texture [4].

Moreover, other counterparts to Fourier phase spectral, such as modules values of wavelet transform, can also reveal some important information such as information of object edges [6, 46].

2.1 Using the phase information by domain filtering

Zhang, et al. (2007) introduced the SR method [18] that is known for its simplicity and very high speed. This method does not use the information of the image color and takes the difference of the logarithm of the image FT and its smoothed image FT, then in order to calculate the saliency map, the FT phase of the input image is used to obtain the inverse FT. Afterwards, HFT method [25] was proposed. The latter method used the color feature with an algorithm similar to SR. In fact, it is the developed version of the SR method. In sum, the SR method calculates the saliency map, using information of the signal FT phase and the specific filtering of signal FT magnitude.

2.2 Using the phase information by normalizing the domain

Zhang et al. (2008) proposed a method known as PFT [17]. This method (unlike SR method) doesn’t use the color feature while having a high speed. The main difference between PFT and SR methods is that the PFT method normalizes FT magnitude (indeed, it considers the strength of all frequencies equal to its unit) and then it takes its reverse FT. In sum, the PFT method detects the salient region by the signal FT phase information.

3 SOD-TS method

SOD-TS algorithm is a task-driven (TD) method using the high level features of task for detecting the salient object. That is, the algorithm detects the salient object with regard to a task that has been specified as an input for the algorithm. But we have to model the task in order to insert it to algorithm as an input.

3.1 Task simulation

Task is a cognitive and qualitative concept whose quantitative simulation is not simple at all because it requires lots of calculative complexities. It can be simulated through some tools such as the deep learning or neural network. In order to avoid the mentioned calculative complexity, in simulating the task in SOD-TS algorithm, we use the geometrical model of the salient object that is to be detected in the image. In fact, this method considers the task as a geometrical shape of the intended salient objects that is to be detected. Then the obtained geometrical model (which represents the feature of the intended salient object to be revealed) can include the binary image; or we want to introduce the intended object to algorithm in a more precise way, it can be a grayscale image.

3.2 How SOD-TS works?

SOD-TS method uses the basic concepts of Fourier Transform that are based on the available information in frequency domain. Before dealing with this method, let recall two important theories on Fourier Transform.

FT phase theory

Changing a signal in the time domain would not lead to the FT magnitude of that signal in the frequency domain; just its phase is changed. This issue can be generalized to the two-dimensional signal (image) such that if a specific object is being changed in the image, its FT magnitude would not change but only its FT phase gets changed.

3.2.1 Conclusion 1: FT phase of any image includes some information about the location of that object in the image

FT absolute theory

If a signal is expanded or contracted (deformed) in the time domain, the subsequent feature in the FT signal is reverse; i.e. if the signal is expanded it gets contracted in the frequency domain, and vice versa. This issue can be generalized to the two-dimensional signal (image) so if the size of a specific object in the image is changed, its FT magnitude is changed as well.

3.2.2 Conclusion 2: FT size of a signal includes information about the geometrical shape of the object in image

Figure 1 shows the structure of SOD-TS algorithm. In this structure, the first input is the original image in which the salient object has to be detected, and the second input is the geometrical model of the object. In the left column of the structure that is related to the original image (first input), first the image is smoothed using Gaussian filter. This operation decreases the overall error of the algorithm because when we smooth the image, the image of its available objects appears more integrated and more uniform and closer to their geometrical model. Then the FT of the filtered image is calculated and its FT phase is extracted. In the right column, the FT of the geometrical model of the object is calculated and its FT magnitude is extracted. Then, using the extracted phase (relating to the FT of the smoothed image of the first input) and the extracted magnitude (relating to the FT of the geometrical model of the object), the reverse FT is calculated resulting in the saliency map. In this obtained saliency map, the points of location of intended objects have the highest rate of lightness. Thus, to detect the salient object, the strongest pixels are selected.
$$ i(m.n)=g(m.n)\ast {i}_0(m.n) $$
(1)
$$ I(u.v)=\mathcal{F}\left\{i(m.n)\right\}={A}_I(x.y){e}^{i{p}_I(x.y)} $$
(2)
$$ M(u.v)=\mathcal{F}\left\{m(m.n)\right\}={A}_M(x.y){e}^{i{p}_M(x.y)} $$
(3)
$$ wmx(m.n)={\mathcal{F}}^{-1}\left\{{A}_M(x.y){e}^{i{p}_I(x.y)}\right\} $$
(4)
$$ \left({m}_i.{n}_i\right)={wmx}^{-1}\left[\max {\left\{ wmx(m.n)\right\}}_i\right]\kern0.75em i=1.2.\dots N $$
(5)
Fig. 1

The structure of algorithm of the proposed method

In equation (1), i0(m. n) is the original image, i(m. n) is the filtered image, and g(m. n) is the Gaussian filter function.

In equation (2), I(u. v) is the FT of the filtered image by Gaussian filter; and AI(x. y) and pI(x. y) are FT size and FT phase, respectively.

In equation (3), M(u. v) is the FT of the model object; and AM(x. y) and pM(x. y) are FT size and FT phase, respectively.

In equation (4), wmx(m. n) is the obtained final image in which the location of strongest pixels implies the location of the object; and.

In equation (5), (mi. ni) denotes the locations of the ith strongest pixels, for i = 1,…,N.

In all of the above equations, * is the convolution operator, \( \mathcal{F} \) and \( {\mathcal{F}}^{-1} \) are the FT operator and inverse FT respectively.

3.3 Justification of the proposed method

Considering the mentioned explanations about the algorithm of the proposed method, we used FT magnitude of the image of object model and the phase of the original image since (based on the conclusion 1) the information of the object location is available in the FT phase of the image. On the other hand, considering the conclusion 2, we use the FT magnitude of the image of object model because the information of the signal appearance is available in the FT magnitude. Moreover, since we look for the location of the object whose appearance is similar to object model, we use the FT phase of the original image that includes the information of location; besides, we use the FT magnitude of image of the object model that includes information of appearance of the object.

3.4 A new justification for phase-only transform (PFT) method

In PFT method, the FT of the image is calculated, then its FT magnitude is considered to be equal to constant signal equal to 1 and then the inverse FT is calculated using this FT magnitude and FT phase of the original image. A closer look at the results of PFT method shows that the mentioned method recognizes the edges and corners very well; or in more general terms, it detects the Diracs (i.e. the points with most difference with their adjacent points) in the image. In the algorithm of this method, the FT magnitude of the image is normalized or is considered to be equal to constant signal equal to 1 that is the FT of Dirac signal. In order to have a better understanding of the subject, we can give a Dirac to the SOD-TS method as an object model. Subsequently, the algorithm will calculate the FT magnitude of the Dirac signal which will be a signal with the constant signal equal to 1 magnitude in all frequencies, and using this FT magnitude and FT phase of the original image, the algorithm calculates the inverse FT where in the obtained image, the points with most similarity to the object model (Dirac signal) have more lightness. These points are the Dirac points, i.e. the points with most difference with their adjacent points (e.g. the edges and corners). This output image is equal to the output of PFT method. In sum, we can conclude that the PFT method is a very specific mode of SOD-TS method (if the Dirac signal is considered as the model in the SOD-TS method, then our method will be the same as the PFT). That is, SOD-TS method is broader and more applicable since it makes it possible to get different results with different models. As shown in Fig. 2, considering the model given to the algorithm of SOD-TS, at a location of the amplitude of the output signal that is more similar to the signal of model, the amplitude signal of the output will have a higher peak.
Fig. 2

(a) Original signal; (b, d,f, h) Signal model; (c, e, g, i) Detection of the intended location

3.5 Advantages of SOD-TS method

3.5.1 Very low processing size

In this method, the task simulation operation and salient object detection are done by simple calculations with low processing size; hence the most important feature of this method is its low processing size and its high speed.

3.5.2 Object detection

Scale-independency and rotation-independency are among the very important characteristics of object detection algorithms. Thus, if an object detection algorithm can meet these characteristics it would be considerably more efficient.

In order to make SOD-TS algorithm independent of scale, the images of the object model are inserted into the algorithm in different scales, and then we select an output of the model that has more adjacent points as the object model. In this regard, we use standard deviation parameter; that is, since each point consists of one length and one width, two quantities of standard deviation can be gained for all points; one for the length of the points and the other for the width of the points. The minimum of these two quantities is selected for the model; and generally the model which has the minimum standard deviation is selected as the suitable model for detecting specified salient object of the image (Fig. 3).
Fig. 3

Scale invarience feature of the object. Object model (g) has more concentraited points and less standard deviation value. (a,d,g,j,m) Object model; (b,e,h,k,n) Saliency map; (c,f,i,l,o) Strongest pixels of the saliency map

On the other hand, for being independent of rotation, we choose the object model in such a way that it does not need any rotation; that is, if the model is rotated, its geometrical shape will not be changed significantly. Even if the object rotates in the image, since its model is independent of rotation, the algorithm will be able to detect the salient abject. In Fig. 4d and g, they have been rotated and since the geometrical shape of the selected model (Fig. 4j) has not been changed significantly by rotation, the algorithm has managed to detect the salient object in the model (Fig. 4f, i).
Fig. 4

Rotation invarience feature using object model (d). (a,d,g) Original image; (b,e,h) Saliency map; (c,f,i) Strongest pixels of the saliency map; (j) Object model

3.5.3 Algorithm controllability in detecting salient object

The object model is one of the parameters which can change the behavior of this algorithm, so that if the object model changes, the salient region will change as well, because as mentioned in section 3.1, SOD-TS algorithm will detect an object as the salient object that is the best response to the current task (object model). This feature is very useful for images consisting multiple salient objects. In Fig. 5a and b there are two salient regions, which have been specified in the images Fig. 5-(e,k) and (h,n) using the models Fig. 5(c,i) and (f,l), respectively.
Fig. 5

Determining the functional domain of the proposed method using object model. (a,b) Original image; (c,f,i,l) Object model; (d,g,j,m) Saliency map; (e,h,k,n) Strongest pixels of the saliency map

It is possible to use this feature for detecting salient regions in images in which the intended object has similar geometrical shapes; and it is possible to use a single model to detect salient regions in which the intended object has the same geometrical shape as the model. As it can be seen in Fig. 6, using a single model (Fig. 6j) the algorithm has detected different ships based on their almost the same geometrical shape. More applications of this algorithm will be presented in the section 4.3. based on the mentioned feature.
Fig. 6

Determining the functional domain of the proposed method using object model (d). (a,d,g) Original image; (b,e,h) Saliency map; (c,f,i) Strongest pixels

4 Results

There are several ways to measure the agreement between model predictions and human annotations [9]. Some metrics evaluate the overlap between a tagged region and model predictions while others try to assess the accuracy of drawn shapes with object boundary. In addition, some metrics have tried to consider both boundary and shape [7, 33]. To evaluate our proposed method, in this article we used two parameters: Area under ROC curve (AUC) and algorithm running time. Moreover, we use the data derived from PASCAL(http://host.robots.ox.ac.uk/pascal/VOC/), EGSSD(http://www.cse.cuhk.edu.hk/leojia/projects/hsaliency/dataset.html), MSRA(https://mmcheng.net/msra10k/).

Receiver operating characteristics (ROC) curve

In addition to the Precision, Recall and , we can also report the false positive rate (FPR) and true positive rate (TPR) when binarizing the saliency map with a set of fixed thresholds:
$$ \mathrm{TPR}=\frac{\mid \mathrm{M}\cap \mathrm{G}\mid }{\mid \mathrm{G}\mid };\mathrm{FRP}=\frac{\mid \overline{\mathrm{M}}\cap \overline{\mathrm{G}}\mid }{\mid \overline{\mathrm{G}}\mid } $$
(6)
where \( \overline{\mathrm{M}} \) and \( \overline{\mathrm{G}} \) denote the complement of the binary mask M and ground-truth, respectively. The ROC curve is the plot of TPR versus FPR by varying the threshold Tf.

Area under ROC curve (AUC) score

While ROC is a two-dimensional representation of a model’s performance, the AUC distills this information into a single scalar. As the name implies, it is calculated as the area under the ROC curve. A perfect model will score an AUC of 1, while random guessing will score an AUC around 0.5.

Table 1 compares SOD-TS method to other methods introduced in [7] and Table 2 compares SOD-TS method to the highest speed method introduced in [7] (i.e. HS [13]) besides two high speed methods of SR and PTF.
Table 1

Comparing the proposed method to other methods in terms of the AUC parameter and running time of the algorithm

#

Model

AUC

1

AC [2]

.565

2

FT [3]

.629

3

CA [16]

.782

4

MSS [1]

.766

5

SEG [35]

.787

6

RS [14]

.713

7

HC [14]

.669

8

SWD [15]

.835

9

SVO [10]

.826

10

CB [19]

.818

11

FES [37]

.854

12

SF [34]

.762

13

LMLC [43]

.800

14

HS [13]

.840

15

GMR [44]

.829

16

DRFI [42]

.897

17

PCA [31]

.848

18

LBI [36]

.828

19

GC [12]

.728

20

CHM [26]

.864

21

DSR [27]

.873

22

MC [20]

.873

23

UFO [21]

.825

24

MNP [32]

.807

25

GR [45]

.794

26

RBD [49]

.867

27

HDCT [22]

.815

28

ST [30]

.868

29

QCUT [5]

.870

30

SOD-TS

.517

Table 2

Comparing the proposed method to other methods in terms of running time of the algorithm

Method

Running time of algorithm (S)

SR [18]

0.2071

PFT [17]

0.0474

HS [13]

0.8173

SOD-TS

0.0640

Figures 7 and 8 compare the proposed method to other methods of salient object detection in image. As the Fig. 1 (relating to AUC parameter) shows, SOD-TS has not have a high precision in the saliency map. But as the Figs. 7 and 8 show, SOD-TS method has managed to detect the salient object with regard to the object model.
Fig. 7

Comparing SOD-TS to other Salient Object Detection methods [7] a. model; b. saliency map; c. strongest pixels

Fig. 8

Comparing SOD-TS to other Salient Object Detection methods [48] a: input; b: HFT [28]; c: BFS [40]; d: BL [38]; e: GL [39]; f: DISC [11]; g: LEGS [41]; h: MDF [24]; i: ELD [23]; j: SOD-LGDRL [48]; k: Ground Truth (GT); l, m, n, o: proposed model (l: object model; m & n: saliency map; o: strongest pixel

4.1 Other examples of SOD-TS

In this section, few examples of SOD-TS are discussed. In each of these examples, the algorithm detects the salient object in the image with the most geometrically similarity to the model with regard to the input geometrical model. Using these examples, we try to present different applications of SOD-TS. Figure 9 is based on Jian li (http://www.escience.cn/people/jianli/DataBase.html) data in which the intended salient object has been detected Fig. 9-(a4, b4, c4, d4, e4, f4, g4) according to the image object model Fig. 9-(a2, b2, c2, d2, e2, f2, g2).
Fig. 9

Samples of the proposed model from jian li data. (a1,b1,c1,d1,e1,f1,g1) Original images; (a2,b2,c2,d2,e2,f2,g2) Object model; (a3,b3,c3,d3,e3,f3,g3) Saliency map; (a4,b4,c4,d4,e4,f4,g4) Strongest pixels of the saliency map

Figure 10 presents the application of the proposed method in optical character recognition (OCR). For this reason, the image of the intended word is considered as the image model and as seen, the algorithm has successfully detected the intended word or letter. Looking carefully, we find that in the input image Fig. 10-(a, e, i, m) the background is black and the words are white while in the model image Fig. 10-(b, f, j, n) it is completely inverse. Nevertheless, the algorithm has been successful in detection. This demonstrates the independence of the algorithm from the brightness level and this is one of the advantages of this method.
Fig. 10

Optical character recognition (OCR). (a,e,I,m) Original images; (b,f,j,n) Object model; (c,g,k,o) Saliency map; (d,h,l,p) Strongest pixels of the saliency map

In Fig. 11a, a rectangular object is hidden inside the blue image. This object cannot be seen clearly because its colure and the pixels around it have the least possible difference with each other. This issue has been illustrated in Fig. 11b, and as seen, the algorithm has well detected this rectangular shaped object according to its model (Fig. 11d).
Fig. 11

Detection of very smoth model. (a) Original images; (b) Original images with object gray level acount; (c) Object model; (d) Strongest pixels of the saliency map; (e) Saliency map

In Fig. 12 the objective is to find the ship and the application of this algorithm has been shown in this field. As known, one of the serious challenges in ship detection is waves in the sea that cause a disturbance as non-target object. It is remarkable that, as seen, this algorithm has successfully detected the ship in Fig. 12d and h, although sea waves are a significant part in these images. Although in Fig. 12i and m unlike Fig. 12a and e sea waves are along the ship and are not separated from that even in this circumstance, using an appropriate model, the algorithm was successful in ship detection (Fig. 12l and p). It is worth mentioning that in Fig. 12m not only the wave is along the boat, but also the intended object is so small that it is fully inside the wave and as seen, the algorithm has successfully detected it.
Fig. 12

Ship detection .(a,e,i,m,q,u) Original image; (b,f,j,n,r,v) Object model; (c,g,k,o,s,w) Saliency map; (d,h,l,p,t,x) Strongest pixels of the saliency map

We encounter another challenge of ship detection when the ship is smoothed in the image. In this situation, the intended ship is smoothed in the image, like Fig. 12q and u. However, using the suitable model, the algorithm of the proposed method has been successful in detecting the ship even in this image (12t, 12x). It should be mentioned that the suitable model to detect the smoothed object in the image is a model which is most similar to the intended object. In fact, the precision of the algorithm is increased by doing so and as seen in Fig. 12r, a model which is very similar to a ship has been selected. The same model is selected for the image in Fig. 12u as well. However, the shape of the ship in this Figure is different from its shape in Fig. 12q.

5 Conclusion

Based on task simulation, in this paper we proposed a salient object detection method that uses the bottom-up attention model. The model proposed in this paper has a high speed and very low processing size due to the use of new features based on FT concepts. Thus it can be easily run for software and hardware implementation. This method has been inspired by the PFT method and it is indeed a very general completion of the PFT method. Moreover, since we simulate the task, this model can have different behaviors with regard to the current task (i.e. object model).

The proposed model does not use the color feature because we believe that the human being manages to detect the salient regions and intended object without relying on the color feature. Since we simulate the task, this model operates with regard to the current task, i.e. the object model because we believe that the concept of task has geometrical components. That is, when a person has a specific task, the intended task can emerge as the geometrical shapes in his mind; and this characteristic is just one of the endless components of the qualitative concept of the task.

Notes

References

  1. 1.
    Achanta R, Susstrunk S (2010) Saliency detection using maximum symmetric surround. IEEE ICIP: 2653–2656Google Scholar
  2. 2.
    Achanta R, Estrada F, Wils P, Susstrunk S (2008) Salient region detection and segmentation. Comp Vis SysGoogle Scholar
  3. 3.
    Achanta R, Hemami S, Estrada F, Susstrunk S (2009) Frequency-tuned salient region detection. IEEE CVPRGoogle Scholar
  4. 4.
    Aiger D, Talbot H (2010) The phase-only transform for unsupervised surface defect detection. Computer Vision and Pattern Recognition (CVPR) IEEE ConferenceGoogle Scholar
  5. 5.
    Aytekin C, Kiranyaz S, Gabbouj M (2014) Automatic object segmentation by quantum cuts. IEEE ICPR: 112–117Google Scholar
  6. 6.
    Bhatnagar G, Wu QMJ, Raman B (2013) Discrete fractional wavelet transform and its application to multiple encryption. Inf Sci 223:297–316MathSciNetCrossRefGoogle Scholar
  7. 7.
    Borji A et al (2015) Salient object detection: a benchmark. IEEE Trans Image Process 24(12)Google Scholar
  8. 8.
    Borji A, Itti L (2013) State-of-the-art in visual attention modeling. IEEE Trans Pattern Anal Mach Intell 35(1)Google Scholar
  9. 9.
    Borji A, Sihite DN, Itti L (2013) What stands out in a scene? A study of human explicit saliency judgment. Vis Res 91(0):62–77CrossRefGoogle Scholar
  10. 10.
    Chang K-Y, Liu T-L, Chen H-T, Lai S-H, (2011) Fusing generic objectness and visual saliency for salient object detection. IEEE ICCV: 914–921Google Scholar
  11. 11.
    Chen T, Lin L, Liu L, Luo X, Li X (2016) Deep image saliency computing via progressive representation learning. IEEE Trans Neural Netw Learn Syst 27(6):1135MathSciNetCrossRefGoogle Scholar
  12. 12.
    Cheng M-M, Warrell J, Lin W-Y, Zheng S, Vineet V, Crook N (2013) Efficient salient region detection with soft image abstraction. IEEE ICCV: 1529–1536Google Scholar
  13. 13.
    Cheng M-M et al. (2015) Global contrast based salient region detection. IEEE TPAMIGoogle Scholar
  14. 14.
    Cheng M-M, Mitra NJ, Huang X, Torr PHS, Hu S-M (2015) Global contrast based salient region detection. IEEE TPAMIGoogle Scholar
  15. 15.
    Duan L, Wu C, Miao J, Qing L, Fu Y (2011) Visual saliency detection by spatially weighted dissimilarity. IEEE CVPR: 473–480Google Scholar
  16. 16.
    Goferman S, Zelnik-Manor L, Tal A (2012) Context-aware saliency detection. IEEE TPAMI 34(10):1915–1926CrossRefGoogle Scholar
  17. 17.
    Gue C, Ma Q, Zhang L (2008) Spatiotemporal saliency detection using phase spectrum of quaternion Fourier transform. Proc IEEE Int Conf Comput Vision Pattern RecognGoogle Scholar
  18. 18.
    Hou X, Zhang L (2007) Saliency detection: a spectral residual approach. Proc IEEE Int Conf Comput Vision Pattern Recogn,: 1–8Google Scholar
  19. 19.
    Jiang H, Wang J, Yuan Z, Liu T, Zheng N, (2011) Automatic salient object segmentation based on context and shape prior. BMVCGoogle Scholar
  20. 20.
    Jiang B, Zhang L, Lu H, Yang C, Yang M-H (2013) Saliency detection via absorbing markov chain. IEEE ICCVGoogle Scholar
  21. 21.
    Jiang P, Ling H, Yu J, Peng J (2013) Salient region detection by ufo: uniqueness, focusness and objectness. IEEE ICCVGoogle Scholar
  22. 22.
    Kim J, Han D, Tai Y-W, Kim J (2014) Salient region detection via high-dimensional color transform. IEEE CVPRGoogle Scholar
  23. 23.
    Lee G, Tai YW, Kim J (2016) Deep saliency with encoded low level distance map and high level features. 2016 IEEE Conf Comput Vision Pattern Recognition (CVPR): 660–668Google Scholar
  24. 24.
    Li G, Yu Y (2015) Visual saliency based on multiscale deep features. Computer Vision and Pattern Recognition: 5455–5463Google Scholar
  25. 25.
    Li J, Levine MD, An X (2007) IEEE Transactions on Pattern Analysis and Machine Intelligence. Class files 6(1)Google Scholar
  26. 26.
    Li X, Li Y, Shen C, Dick AR, van den Hengel A (2013) Contextual hypergraph modeling for salient object detection. IEEE ICCV: 3328–3335Google Scholar
  27. 27.
    Li X, Lu H, Zhang L, Ruan X, Yang M-H (2013) Saliency detection via dense and sparse reconstruction. IEEE ICCVGoogle Scholar
  28. 28.
    Li J, Levine MD, An X, Xu X, He H (2013) Visual saliency based on scale-space analysis in the frequency domain. IEEE Trans Pattern Anal Mach Intell 35(4):996–1010CrossRefGoogle Scholar
  29. 29.
    Liu N, Han J (2018) A deep spatial contextual long-term recurrent convolutional network for saliency detection. IEEE Trans. Image ProcessGoogle Scholar
  30. 30.
    Liu Z, Zou W, Le Meur O (2013) Saliency tree: a novel saliency detection framework. IEEE TIPGoogle Scholar
  31. 31.
    Margolin R, Tal A, Zelnik-Manor L (2013) What makes a patch distinct?. IEEE CVPR: 1139–1146Google Scholar
  32. 32.
    Margolin R, Zelnik-Manor L, Tal A (2013) Saliency for image manipulation. Vis Comput: 1–12Google Scholar
  33. 33.
    Movahedi V, Elder JH (2010) Design and perceptual validation of performance measures for salient object segmentation. IEEE CVPRWGoogle Scholar
  34. 34.
    Perazzi F, Krahenbuhl P, Pritch Y, Hornung A (2012) Saliency filters: contrast based filtering for salient region detection. IEEE CVPR: 733–740Google Scholar
  35. 35.
    Rahtu E, Kannala J, Salo M, Heikkila J (2010) Segmenting salient objects from images and videos. ECCVGoogle Scholar
  36. 36.
    Siva P, Russell C, Xiang T, Agapito L (2013) Looking beyond the image: unsupervised learning for object saliency and detection. IEEE CVPR: 3238–3245Google Scholar
  37. 37.
    Tavakoli HR, Rahtu E, Heikkila J (2011) Fast and efficient saliency detection using sparse sampling and kernel density estimation. Scandinavian conference on image analysis: 666–675Google Scholar
  38. 38.
    Tong N, Lu H, Ruan X, Yang M-H (2015) Salient object detection via bootstrap learning. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition: 1884–1892Google Scholar
  39. 39.
    Tong N, Lu H, Zhang Y, Ruan X (2015) Salient object detection via global and local cues. Pattern Recogn 48(10):3258–3267CrossRefGoogle Scholar
  40. 40.
    Wang J, Lu H, Li X, Tong N, Liu W (2015) Saliency detection via background and foreground seed selection. Neurocomputing 152:359–368CrossRefGoogle Scholar
  41. 41.
    Wang L, Lu H, Xiang R, Yang MH (2015) Deep networks for saliency detection via local estimation and global search. Comput Vision Pattern Recogn: 3183–3192Google Scholar
  42. 42.
    Wang J, Jiang H, Yuan Z, Cheng M-M, Hu X, Zheng N (2017) Salient object detection: a discriminative regional feature integration approach. Int J Comput Vis 123(2):251–268CrossRefGoogle Scholar
  43. 43.
    Xie Y, Lu H, Yang M-H (2013) Bayesian saliency via low and mid-level cues. IEEE TIP 22(5)Google Scholar
  44. 44.
    Yang C, Zhang L, Lu H, Ruan X, Yang M-H (2013) Saliency detection via graph-based manifold ranking. IEEE CVPRGoogle Scholar
  45. 45.
    Yang C, Zhang L, Lu H (2013) Graph-regularized saliency detection with convex-hull-based center prior. IEEE Signal Process Lett 20(7):637–640CrossRefGoogle Scholar
  46. 46.
    You X, Du L, Cheung Y-m, Chen Q (2010) A blind watermarking scheme using new nontensor product wavelet filter banks. IEEE Trans Image Process 19(12):3271–3284MathSciNetCrossRefGoogle Scholar
  47. 47.
    Zhang J et al (2018) Top-down neural attention by excitation backprop. Int J Comput Vis 126:1084–1102CrossRefGoogle Scholar
  48. 48.
    Zhu D (2018) Salient object detection via a local and global method based on deep residual network. J Vis Commun Image Represent 54Google Scholar
  49. 49.
    Zhu W, Liang S, Wei Y, Sun J (2014) Saliency optimization from robust background detection. IEEE CVPRGoogle Scholar

Copyright information

© The Author(s) 2019

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (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.

Authors and Affiliations

  1. 1.Khorasan Institute of Higher EducationMashhadIran
  2. 2.Faculty of EngineeringFerdowsi University of MashhadMashhadIran

Personalised recommendations