Saliency Detection: A Spectral Residual Approach

By Xiaodi Hou, Liqing Zhang. 2007 IEEE Conference on Computer Vision and Pattern Recognition, 2007

3,643 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/9ce3c18eb4fa86fd19bce46227be39895de4e4ab

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cognitive Neuroscience

In Computer Vision and Pattern Recognition

Citations in the corpus

Measuring the Objectness of Image Windows (2011, methodology): “We incorporate a state-of-the-art saliency detector [27] as one cue into our objectness measure.”
Saliency Detection via Absorbing Markov Chain (2013, neutral): “SED: On this single object and two object dataset, we compare the proposed method with eleven state-of-the-art methods which are LR [29], CB [17], SVO [7], RC [8], CA [27], SER [31], FT [2], GB [14], SR [15], LC [37] and IT [16].”
Fusing generic objectness and visual saliency for salient object detection (2011, background): “Visual attention relates to how humans process different information in a scene, and is primarily analyzed through investigating the allocation of eye fixations.”
Revisiting Video Saliency: A Large-Scale Benchmark and a New Model (2018, background): “Early static saliency models [36, 69, 15, 7, 18, 22, 33, 63] are mostly based on the contrast assumption that conspicuous visual features “pop-out” and involuntarily capture attention (see [5, 6] for review).”
Spatio-temporal Saliency detection using phase spectrum of quaternion fourier transform (2008, methodology): “After careful analysis, we find that SR of the amplitude spectrum is not essential to obtain the saliency map in [10]; however, the saliency map can be calculated by the image’s Phase spectrum of Fourier Transform (PFT) alone.”
BING: Binarized normed gradients for objectness estimation at 300fps (2014, neutral): “The 3 methods [3, 48, 57] have been evaluated on the same benchmark and shown to outperform other alternative proposal methods [6,21,25,30,50], saliency measures [33, 36], interesting point detectors [44], and HOG detector [17] (see [3] for the comparisons).”
Two-Stage Learning to Predict Human Eye Fixations via SDAEs (2016, methodology): “However, as pointed out by Zhang et al. [36], strong center-bias may exist in human fixations and it can influence the performance of a saliency model.”
Visual saliency detection by spatially weighted dissimilarity (2011, methodology): “Actually, the method in [15] was based on the well-known 1/ f law (i.e., scale invariance) which measures the spatial correlations between pixels in an image [17].”
Hierarchical Image Saliency Detection on Extended CSSD (2016, methodology): “AC [10] CA [31] FT [29] GB [21] HC [1] IT [19] LC [28] LR [30] SR [47] RC [1] RCC [1] HS [16] CHS”
Advanced Deep-Learning Techniques for Salient and Category-Specific Object Detection: A Survey (2018, background): “It first obtained an initial set of proposals from the salient locations [32] in a given image.”
Automatic salient object segmentation based on context and shape prior (2011, methodology): “And we compared our approach with other state-of-the-art methods [1, 7, 11, 12, 13].”
AutoFocus: Efficient Multi-Scale Inference (2018, background): “AutoFocus shares the motivation with saliency based methods which perform a guided search while processing images [26, 24, 39, 48], but it is designed to predict small objects in coarser scales and they need not be salient.”
Perceptual visual quality metrics: A survey (2011, background): “The VA map can be determined as [58]:”
A unified approach to salient object detection via low rank matrix recovery (2012, methodology): “Compared with IT[12], MZ[18], IG[1], RC[4], GB[9], SR[10] and CA[8], our method achieves the best performance.”
Deep Learning for Time Series Anomaly Detection: A Survey (2022, positive): “[144] combine the Spectral Residual (SR) model with CNN (SR-CNN) to achieve superior accuracy, as the SR model derived from visual saliency detection [82].”
Geodesic Saliency Using Background Priors (2012, methodology): “GS GD GS SP IT [1] SR [21] GB [22] FT [9] CA [11] LC [33] HC [12] RC [12]”
An Assisted Photography Framework to Help Visually Impaired Users Properly Aim a Camera (2014, methodology): “Different situations call for different solutions, and so we chose to test several methods with a variety of street photographs [Walther and Koch 2006; Hou and Zhang 2007; Guo et al. 2008; Bian and Zhang 2008; Achanta et al. 2009].”
A deep multi-level network for saliency prediction (2016, methodology): “Inspired by biological studies, researchers have deﬁned hand-crafted and multi-scale features that capture a large spectrum of stimuli: lower-level features (color, texture, contrast) [6] or higher-level concepts (faces, people, text, horizon) [7].”
Salient Region Detection by UFO: Uniqueness, Focusness and Objectness (2013, methodology): “Also some methods turn to the frequency domain to search for saliency cues [10, 13, 21].”