Modeling Visual Clutter Perception using Proto-Object Segmentation

Yu, Chen-Ping; Samaras, Dimitris; Zelinsky, Gregory J.

doi:10.1167/14.10.365

Cited by 11 publications

(27 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a natural scene, however, an “object” is a perceptual and hierarchical construct that can change depending on the task, context, and mindset of the observer. This renders quantifying the number of items in a natural scene difficult, and some even argue that the problem of object segmentation is ill‐conceived . There are two main strategies for experimentally relating fixation patterns to objects in a scene.…”

Section: Discussionmentioning

confidence: 99%

“…First, one can ask observers to label the scene and relate this ground truth to fixation patterns . Second, one can delineate preattentive objects based on features that can be computed from the stimulus . The resulting entities, which can potentially gain objecthood, are frequently identified as proto‐objects .…”

Section: Discussionmentioning

confidence: 99%

“…The resulting entities, which can potentially gain objecthood, are frequently identified as proto‐objects . The conditions under which such proto‐objects behave like real objects for fixation selection is an interesting issue in itself . In this study, we restricted ourselves to features that can be computed from the current stimulus and refrain from relating them to any semantic scene content or context.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

A new approach to modeling the influence of image features on fixation selection in scenes

Nuthmann

Einhäuser

2015

Annals of the New York Academy of Sciences

View full text Add to dashboard Cite

Which image characteristics predict where people fixate when memorizing natural images? To answer this question, we introduce a new analysis approach that combines a novel scene-patch analysis with generalized linear mixed models (GLMMs). Our method allows for (1) directly describing the relationship between continuous feature value and fixation probability, and (2) assessing each feature's unique contribution to fixation selection. To demonstrate this method, we estimated the relative contribution of various image features to fixation selection: luminance and luminance contrast (low-level features); edge density (a mid-level feature); visual clutter and image segmentation to approximate local object density in the scene (higher-level features). An additional predictor captured the central bias of fixation. The GLMM results revealed that edge density, clutter, and the number of homogenous segments in a patch can independently predict whether image patches are fixated or not. Importantly, neither luminance nor contrast had an independent effect above and beyond what could be accounted for by the other predictors. Since the parcellation of the scene and the selection of features can be tailored to the specific research question, our approach allows for assessing the interplay of various factors relevant for fixation selection in scenes in a powerful and flexible manner.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

A new approach to modeling the influence of image features on fixation selection in scenes

Nuthmann

Einhäuser

2015

Annals of the New York Academy of Sciences

View full text Add to dashboard Cite

show abstract

“…Bravo and Farid's Scale Invariant clutter metric [194] was designed to predict the performance (mean search time) of observers searching for non-salient targets with unknown features at unknown positions in unstructured scenes and may therefore especially applicable to camouflaged targets. Another clutter metric that is directly associated with the amount of detail in a scene is the Number of Proto-Objects (NPO) or the number of image segments or super-pixels with similar intensity, color and gradient orientation features (the proto-objects: [195]). Larger NPO values correspond to higher levels of visual clutter.…”

Section: Global Clutter Metricsmentioning

confidence: 99%

Review of camouflage assessment techniques

Toet

Hogervorst

2020

Target and Background Signatures VI

View full text Add to dashboard Cite

In military operations signature reduction techniques such as camouflage nets, low-emissive paints, and camouflage patterns are typically deployed to optimize the survivability of high value assets by minimizing their detectability. Various methods have been developed to assess the effectiveness of these camouflage measures. There are three main approaches to the evaluation of camouflage measures: (1) a subjective approach through observer experiments, (2) an objective computational approach through image analysis, and (3) an objective approach through physical measurements. Although subjective evaluation methods have a direct relation with the operational practice, they are often difficult to implement because of time and budget restrictions (these experiments typically require a lot of observers, repetitions and different targets and backgrounds), or simply because the associated conditions are not safe for the observers. The objective evaluation methods are typically based on the outcome of psychophysical laboratory experiments using simple artificial stimuli, presented under extremely restricted (impoverished) conditions, and in different experimental paradigms. Other objective methods are based on signal processing techniques and have no obvious counterpart in human vision. Only a few attempts have been made to validate any of these objective metrics against the performance of human observers in realistic military scenarios. As a result, there are currently no standard and internationally accepted methods and procedures to evaluate camouflage equipment and techniques, and to indicate their military effectiveness. In this review paper we present an overview of the various subjective (psychophysical) and objective (computational, image or video based) evaluation methods that are currently available and that have been used to validate camouflage effectiveness. In addition, we discuss the relative merits of field experiments versus laboratory experiments.

show abstract

“…Yu et al [22] proposed a proto-object model of visual clutter. They note that good progress has been made at quantifying clutter by counting objects, giving an estimate of set size, and observe that image segmentations are imperfect ground truths for clutter because they are subjective.…”

Section: Complexity and Visual Cluttermentioning

confidence: 99%

Predicting human complexity perception of real-world scenes

Nagle

Lavie

2020

R. Soc. open sci.

View full text Add to dashboard Cite

Perceptual load is a well-established determinant of attentional engagement in a task. So far, perceptual load has typically been manipulated by increasing either the number of task-relevant items or the perceptual processing demand (e.g. conjunction versus feature tasks). The tasks used often involved rather simple visual displays (e.g. letters or single objects). How can perceptual load be operationalized for richer, real-world images? A promising proxy is the visual complexity of an image. However, current predictive models for visual complexity have limited applicability to diverse real-world images. Here we modelled visual complexity using a deep convolutional neural network (CNN) trained to learn perceived ratings of visual complexity. We presented 53 observers with 4000 images from the PASCAL VOC dataset, obtaining 75 020 2-alternative forced choice paired comparisons across observers. Image visual complexity scores were obtained using the TrueSkill algorithm. A CNN with weights pre-trained on an object recognition task predicted complexity ratings with r = 0.83. By contrast, feature-based models used in the literature, working on image statistics such as entropy, edge density and JPEG compression ratio, only achieved r = 0.70. Thus, our model offers a promising method to quantify the perceptual load of real-world scenes through visual complexity.

show abstract

Modeling Visual Clutter Perception using Proto-Object Segmentation

Cited by 11 publications

References 0 publications

A new approach to modeling the influence of image features on fixation selection in scenes

A new approach to modeling the influence of image features on fixation selection in scenes

Review of camouflage assessment techniques

Predicting human complexity perception of real-world scenes

Contact Info

Product

Resources

About