Visual saliency detection based on homology similarity and an experimental evaluation

Saliency mechanism has been considered crucial in the human visual system and helpful to object detection and recognition. This paper addresses an information theoretic model for visual saliency detection. It consists of two steps: first, using the Non-negative Matrix Factorization with sparseness constraints (NMFsc) algorithm to learn the basis functions from a set of randomly sampled natural image patches; and then, applying information theoretic principle to generate the saliency map by the Salient Information (SI) which is calculated from the coefficients represented by basis functions. We compare our model with the previous methods on natural images. Experimental results show that our model performs better than existing approaches.

Visual saliency detection has become a challenging area in computer vision. In this paper, we propose a novel region based saliency detection model which considers background priors. The proposed method consists of two successive steps - region weighting and contrast computing. In the step of region weighting, we calculate the region weight for each region by region-level image feature and a log-linear prediction model. In the step of contrast computing, we propose a modified contrast computing algorithm by exploiting the advantage of region weights for bottom-up saliency detection. The experimental results on two datasets prove that our method effectively improves the performance on visual saliency detection.

Due to complex lighting circumstances and other factors, visual saliency recognition of underwater images is not as good as visual saliency detection models that function well on land when applied to underwater environments. The dearth of large-scale annotated underwater picture datasets has further hindered the development of visual saliency detection in underwater images. we propose an attention-based approach for underwater image saliency detection to address this issue. The backbone network is pre-trained using both land and underwater image datasets, which effectively solves the problem of fewer datasets. After that, a residual refinement module incorporating a channel-space attention mechanism is merged, the model as a whole is trained end-to-end using underwater image datasets, and the model accuracy is experimentally tested. The experimental results in three underwater image datasets show that our method is better than several other methods in the mainstream evaluation indicators of the saliency detection.

The effectiveness of depth information in saliency detection has been fully proved. However, it is still worth exploring how to utilize the depth information more efficiently. Erroneous depth information may cause detection failure, while non-salient objects may be closer to the camera which also leads to erroneously emphasis on non-salient regions. Moreover, most of the existing RGB-D saliency detection models have poor robustness when the salient object touches the image boundaries. To mitigate these problems, we propose a multi-stage saliency detection model with the bilateral absorbing Markov chain guided by depth information. The proposed model progressively extracts the saliency cues with three level (low-, mid-, and high-level) stages. First, we generate low-level saliency cues by explicitly combining color and depth information. Then, we design a bilateral absorbing Markov chain to calculate mid-level saliency maps. In mid-level, to suppress boundary touch problem, we present the background seed screening mechanism (BSSM) for improving the construction of the two-layer sparse graph and better selecting background-based absorbing nodes. Furthermore, the cross-modal multi-graph learning model (CMLM) is designed to fully explore the intrinsic complementary relationship between color and depth information. Finally, to obtain a more highlighted and homogeneous saliency map in high-level, we structure a depth-guided optimization module which combines cellular automata and suppression-enhancement function pair. This optimization module refines the saliency map in color space and depth space, respectively. Comprehensive experiments on three challenging benchmark datasets demonstrate the effectiveness of our proposed method both qualitatively and quantitatively.

Even though there have been great advancements in computer vision tasks, the development of human visual attention models is still not well investigated. In day-to-day life, one can find ample applications of saliency detection in image and video processing. This paper presents an efficient visual saliency detection model based on Ripplet transform, which aims at detecting the salient region and achieving higher Receiver Operating Characteristics (ROC). Initially the feature maps are obtained from Ripplet transform in different scales and different directions of the image. The global and local saliency maps are computed based on the global probability density distribution and feature distribution of local areas, which are combined together to get the final saliency map. Ripplet-transform-based visual saliency detection is the novel approach carried out in this paper. Experimental results indicate that the proposed method based on Ripplet transformation can give excellent performance in terms of precision, recall, F measure and Mean Absolute Error (MAE), and is compared with 10 state-of-the-art methods on five benchmark datasets.

A novel visual saliency detection method for infrared video sequences

Visual saliency detection plays a significant role in the fields of computer vision. In this paper, we introduce a novel saliency detection method based on weighted linear multiple kernel learning (WLMKL) framework, which is able to adaptively combine different contrast measurements in a supervised manner. As most influential factor is contrast operation in bottom-up visual saliency, an average weighted corner-surround contrast (AWCSC) is first designed to measure local visual saliency. Combined with common-used center-surrounding contrast (CESC) and global contrast (GC), three types of contrast operations are fed into our WLMKL framework to produce the final saliency map. We show that the assigned weights for each contrast feature maps are always normalized in our WLMKL formulation. In addition, the proposed approach benefits from the advantages of the contribution of each individual contrast feature maps, yielding more robust and accurate saliency maps. We evaluated our method for two main visual saliency detection tasks: human fixed eye prediction and salient object detection. The extensive experimental results show the effectiveness of the proposed model, and demonstrate the integration is superior than individual subcomponent.

Object region proposal has a wide application on various domains, such as object detection, object tracking, robot navigation, and anomaly detection. Widely used region proposal methods are based on either grouping superpixels or sliding windows. Previous studies have been done by grouping either convolution features, superpixels, or windows. Moreover, histogram of gradients (HoGs) of lower-level features (pixels) can be used for the region proposal task. In this study, both the concepts of HoGs and convolution features are used for proposing object(s) region(s) in an image. Region proposal is done by HoGs of convolution features and statistical application on the gradient data of different feature maps. First, a sized image is passed through one convolution layer to generate reduced feature maps. Then, a window slides over each feature map and HoG is applied over each window. Thereafter, the histogram is analysed using the concept of Coefficient of Variation (CV) and angle (slope) applied on histogram data values. A window is considered as the object region if the CV and angle of its histogram data exceed some thresholds. For all feature maps, we have obtained several widows that represent approximated object regions. These windows are grouped to obtain the possible region(s) proposal. The effectiveness of the proposed method is demonstrated over one of the benchmark datasets, called `PASCAL VOC'. Our method is found superior to two other state-of-the-art, namely region-based segmentation (RBS), and combining efficient object localization and image classification (CEOLIC) in terms of recall.

Biologically inspired visual object tracking has been very popular in the last decades. Several methods have exploited some visual cognitive mechanisms to outperform trackers purely based on machine vision in accuracy and speed. Although they have shown outstanding performance on recent benchmarks, they are still far from achieving the performance of a human in tracking. Visual attention is an important mechanism, which can be applied by saliency detection models. Saliency detection models can be determined in the spatial domain or the frequency domain. Spatial domain based models mostly have complex computations and are time-consuming. Thus, we have improved and recruited Frequency Divisive Normalization (FDN) model as a saliency detection method to propose a bio-inspired object tracking algorithm. This algorithm has been developed based on Fast Fourier Transform (FFT) and uses spectral features in order to achieve a high-speed tracker. It also can be launched on normal hardware. FDN model is based on Feature Integration Theory (FIT) and operates all the processes of saliency detection in the frequency domain, including decomposing the image to the features, lateral inhibition interaction, combining all features, and computing the saliency map. In addition to, extracting color, intensity and orientation it uses spatial frequency features, which is rarely used by other trackers. Extensive evaluations on large-scale benchmark datasets show that the proposed method has good performance with respect to its runtime.

This study aims to improve the quality of multi-view video, and designs a new method by integrating Convolutional neural network (CNN), visual saliency detection and image enhancement theory. The experimental results show that the proposed visual saliency detection model and convolution filter sensor have made remarkable progress. The superiority of the visual saliency detection model is helpful to accurately locate the key features of the image and provide accurate targets for subsequent enhancement processing. The convolution filter sensor improves the peak value of the image, narrows the gap with the original image and improves the visual effect. Supplementary experiments further verify the effectiveness of the method. Through the quantitative comparison between SSIM and MS-SSIM, the method is obviously superior to the existing methods on several data sets, showing a robust video quality enhancement effect. These results highlight the superiority and robustness of the method, and bring strong empirical support to the field of multi-view video quality enhancement, which is expected to have an important impact in practical applications.

Deep (machine) learning in recent years has significantly increased the predictive modeling strengths for applications in the areas of computer vision, speech recognition, and natural language processing. Being heavily data-driven, conventional general-purpose deep learning methods mainly rely on training data to learn task-specific data representations. In this thesis, we focus on leveraging novel learning algorithms in deep learning to tackle visual recognition challenges at the three key levels -- namely pixel, object, and image levels. Throughout the thesis, we demonstrate how each and every of our proposed methods can address a specific visual recognition problem. Firstly, we apply deep learning to improve the saliency detection performance of deep networks. A coarse saliency map of the input image's salient object is initially predicted by a convolutional-deconvolutional network (or equivalently a fully convolutional network). Following that, the coarse saliency map is sequentially refined by a separate network called Recurrent Attentional Convolutional-Deconvolutional Network (RACDNN) that shares its network weights across its sequential steps. At every sequential (temporal) step, the same RACDNN is applied to a newly selected/attended image sub-region and it performs saliency map refinement for that particular sub-region. Despite having many learnable network parameters and no access to additional training data, RADCNN can be trained well and it outperforms the "single-step" (coarse saliency detection) baseline and state-of-the-art saliency detection methods by large margins, on several saliency detection benchmark datasets. Secondly, we introduce auto-encoder weight transfer networks (AE-WTN+), a deep model for scaling up object detection. Conventional WTN transfers object class knowledge from the classification weights of a pretrained large-scale image classification network, to the classification weights of an object detection network. However, they suffer from under-fitting and are not able to preserve the rich object class information of novel classes. In this work, we propose a new WTN method -- AE-WTN+ which makes use of normalization techniques to ease network optimization. Inputs and intermediate features are normalized to improve trainability for reduced under-fitting. More importantly, a novel auto-encoding (reconstruction) loss is proposed to encourage information preservation of all classes in the weight predictions of AE-WTN+. This enables the predicted weights of AE-WTN+ to maintain the class discriminability and neighbourhood relationships contained in the pre-trained large-scale classification weights. Experiments on large-scale detection datasets validate the effectiveness of AE-WTN+ on both seen and novel classes. Thirdly, we improve the efficiency of cross-layer network connections for very deep convolutional network in the context of image classification. This gives rise to a new convolutional network architecture called Deluge Networks (DelugeNets) which pass the output (feature activations) of any preceding composite network layer as direct inputs to all succeeding composite layers. For the incoming feature activations of preceding composite layers, the succeeding composite layers first apply rather lightweight cross-layer depthwise convolutional weights on them for feature aggregation. The heavy lifting is carried out by the last convolutional weight layer in each preceding composite layer that performs cross-channel convolutions. The proposed technique enables unimpeded information flow among many layers in very deep DelugeNets with great efficiency. Experiments on small-scale and large-scale datasets show that DelugeNets perform competitively to state-of-the-art deep convolutional networks at lower computational and parameter costs. Fourthly, we propose Stochastic Downsampling (SDPoint), a method to overcome a major limitation suffered by conventional deep convolutional networks -- their computational costs during inference are fixed and cannot be changed. SDPoint downsamples the feature maps at a randomly selected point (layer index) in the network hierarchy during training. The many downsampling configurations known as SDPoint instances each entail a unique downsampling point and downsampling ratio. The same convolutional network weights are shared across the different SDPoint instances that downsample feature maps differently. During inference, the computational cost of a network trained with SDPoint can be instantaneously adjusted to fit a given computational budget, by manually selecting an appropriate SDPoint instance. Additionally, sharing weights between different feature map scales provides significant regularization benefits, making the networks trained with SDPoint less sensitive to scales and more competent. In image classification experiments, SDPoint demonstrates significant cost-accuracy and computational advantages, over independently-trained (non-weight-shared) baselines with various inference costs.

Global feature integration based salient region detection

Since 20th century generation, image segmentation has been attached great importance by the people and thousands of the segmentation algorithms have been proposed. In practice, in order to facilitate research and analysis of the images that often only needs to be interested in certain parts of the image part is divided into a certain properties of a particular area, in order to use the target further, we also need to complete for their separation. This paper integrates the visual attention model to propose the novel image segmentation algorithm based on visual saliency detection and the integrated feature extraction. We have achieved the better numerical performance compared with other state-of-the-art methodologies. The propose method integrates the local and the regional features to construct the global optimized saliency map to serve as the basis for segmentation. The entropy of the image is used for determining the threshold and the effectiveness is proved through experiment.

Data-driven saliency detection has attracted strong interest as a result of applying convolutional neural networks to the detection of eye fixations. Although a number of image-based salient object and fixation detection models have been proposed, video fixation detection still requires more exploration. Different from image analysis, motion and temporal information is a crucial factor affecting human attention when viewing video sequences. Although existing models based on local contrast and low-level features have been extensively researched, they failed to simultaneously consider interframe motion and temporal information across neighboring video frames, leading to unsatisfactory performance when handling complex scenes. To this end, we propose a novel and efficient video eye fixation detection model to improve the saliency detection performance. By simulating the memory mechanism and visual attention mechanism of human beings when watching a video, we propose a step-gained fully convolutional network by combining the memory information on the time axis with the motion information on the space axis while storing the saliency information of the current frame. The model is obtained through hierarchical training, which ensures the accuracy of the detection. Extensive experiments in comparison with 11 state-of-the-art methods are carried out, and the results show that our proposed model outperforms all 11 methods across a number of publicly available datasets.

In this paper, a novel and efficient framework by exploiting Quaternionic Distance Based Weber Local Descriptor (QDWLD) and object cues is proposed for image saliency detection. In contrast to the existing saliency detection models, the advantage of the proposed approach is that it can combine quaternion number system and object cues simultaneously, which is independent of image contents and scenes. Firstly, QDWLD, which was initially designed for detecting outliers in color images, is used to represent the directional cues in an image. Meanwhile, two low-level priors, namely the Convex-Hull-Based center and color contrast cue of the image, are utilized and fused as an object-level cue. Finally, by combining QDWLD with object cues, a reliable saliency map of the image can be computed and estimated. Experimental results, based on two widely used and openly available database, show that the proposed method is able to produce reliable and promising salient maps/estimations, compared to other state-of-the-art saliency-detection models.