Saliency Detection Technique Research Articles

Automatic Optic Disc Boundary Extraction Based on Saliency Object Detection and Modified Local Intensity Clustering Model in Retinal Images

Accurate optic disc localization and segmentation are two main steps when designing automated screening systems for diabetic retinopathy. In this paper, a novel optic disc detection approach based on saliency object detection and modified local intensity clustering model is proposed. It consists of two stages: in the first stage, the saliency detection technique is introduced to the enhanced retinal image with the aim of locating the optic disc. In the second stage, the optic disc boundary is extracted by the modified Local Intensity Clustering (LIC) model with oval-shaped constrain. The performance of our proposed approach is tested on the public DIARETDB1 database. Compared to the state-of-the-art approaches, the experimental results show the advantages and effectiveness of the proposed approach.

Saliency-based abnormal event detection in crowded scenes

Abnormal event detection plays a critical role for intelligent video surveillance, and detection in crowded scenes is a challenging but more practical task. We present an abnormal event detection method for crowded video. Region-wise modeling is proposed to address the inconsistent detected motion of the same object due to different depths of field. Comparing to traditional block-wise modeling, the region-wise method not only can reduce heavily the number of models to be built but also can enrich the samples for training the normal events model. In order to reduce the computational burden and make the region-based anomaly detection feasible, a saliency detection technique is adopted in this paper. By identifying the salient parts of the image sequences, the irrelevant blocks are ignored, which removes the disturbance and improves the detection performance further. Experiments on the benchmark dataset and comparisons with the state-of-the-art algorithms validate the advantages of the proposed method.

Perception-oriented video saliency detection via spatio-temporal attention analysis

Human visual system actively seeks salient regions and movements in video sequences to reduce the search effort. Computational visual saliency detection model provides important information for semantic understanding in many real world applications. In this paper, we propose a novel perception-oriented video saliency detection model to detect the attended regions for both interesting objects and dominant motions in video sequences. Based on the visual orientation inhomogeneity of human perception, a novel spatial saliency detection technique called visual orientation inhomogeneous saliency model is proposed. In temporal saliency detection, a novel optical flow model is created based on the dynamic consistency of motion. We fused the spatial and the temporal saliency maps together to build the spatio-temporal attention analysis model toward a uniform framework. The proposed model is evaluated on three typical video datasets with six visual saliency detection algorithms and achieves remarkable performance. Empirical validations demonstrate the salient regions detected by the proposed model highlight the dominant and interesting objects effectively and efficiently. More importantly, the saliency regions detected by the proposed model are consistent with human subjective eye tracking data.

Visual Saliency Detection With Free Energy Theory

Visual saliency can be thought of as the product of human brain activity. Most existing models were built upon local features or global features or both. Lately, a so-called free energy principle unifies several brain theories within one framework, and tells where easily surprise human viewers in a visual stimulus through a psychological measure. We believe that this “surprise” should be highly related to visual saliency, and thereby introduce a novel computational Free Energy inspired Saliency detection technique (FES). Our method computes the local entropy of the gap between an input image signal and its predicted counterpart that is reconstructed from the input one with a semi-parametric model. Experimental results prove that our algorithm predicts human fixation points accurately and is superior to classical/state-of-the-art competitors.

Flame detection algorithm based on a saliency detection technique and the uniform local binary pattern in the YCbCr color space

Computer vision-based fire detection involves flame detection and smoke detection. This paper proposes a new flame detection algorithm that is based on a saliency detection technique and on the uniform local binary pattern (ULBP). In still images and video sequences, an area that contains an open flame is always noticeable because fire is an exceptional event. Thus, to utilize the color information of flame pixels, the probability density function (pdf) of the flame pixel color can be obtained using Parzen window nonparametric estimation. This a priori pdf is then fused with the saliency detection phase as top-down information so that the flame candidate area can be extracted. To reduce the number of false alarms, the image texture of the candidate area is analyzed by ULBP, and an exponential function with two parameters is utilized to model the texture of the flame area. According to the experimental results, our proposed method can reduce the number of false alarms greatly compared with an alternative algorithm, while ensuring the accurate classification of positive samples. The classification performance of our proposed method is proven to be better than that of alternative algorithms.

A Co-Saliency Model of Image Pairs

In this paper, we introduce a method to detect co-saliency from an image pair that may have some objects in common. The co-saliency is modeled as a linear combination of the single-image saliency map (SISM) and the multi-image saliency map (MISM). The first term is designed to describe the local attention, which is computed by using three saliency detection techniques available in literature. To compute the MISM, a co-multilayer graph is constructed by dividing the image pair into a spatial pyramid representation. Each node in the graph is described by two types of visual descriptors, which are extracted from a representation of some aspects of local appearance, e.g., color and texture properties. In order to evaluate the similarity between two nodes, we employ a normalized single-pair SimRank algorithm to compute the similarity score. Experimental evaluation on a number of image pairs demonstrates the good performance of the proposed method on the co-saliency detection task.