Effective Set Size Research Articles

Patch-based generative adversarial neural network models for head and neck MR-only planning.

To evaluate pix2pix and CycleGAN and to assess the effects of multiple combination strategies on accuracy for patch-based synthetic computed tomography (sCT) generation for magnetic resonance (MR)-only treatment planning in head and neck (HN) cancer patients. Twenty-three deformably registered pairs of CT and mDixon FFE MR datasets from HN cancer patients treated at our institution were retrospectively analyzed to evaluate patch-based sCT accuracy via the pix2pix and CycleGAN models. To test effects of overlapping sCT patches on estimations, we (a) trained the models for three orthogonal views to observe the effects of spatial context, (b) we increased effective set size by using per-epoch data augmentation, and (c) we evaluated the performance of three different approaches for combining overlapping Hounsfield unit (HU) estimations for varied patch overlap parameters. Twelve of twenty-three cases corresponded to a curated dataset previously used for atlas-based sCT generation and were used for training with leave-two-out cross-validation. Eight cases were used for independent testing and included previously unseen image features such as fused vertebrae, a small protruding bone, and tumors large enough to deform normal body contours. We analyzed the impact of MR image preprocessing including histogram standardization and intensity clipping on sCT generation accuracy. Effects of mDixon contrast (in-phase vs water) differences were tested with three additional cases. The sCT generation accuracy was evaluated using mean absolute error (MAE) and mean error (ME) in HU between the plan CT and sCT images. Dosimetric accuracy was evaluated for all clinically relevant structures in the independent testing set and digitally reconstructed radiographs (DRRs) were evaluated with respect to the plan CT images. The cross-validated MAEs for the whole-HN region using pix2pix and CycleGAN were 66.9±7.3 vs 82.3±6.4HU, respectively. On the independent testing set with additional artifacts and previously unseen image features, whole-HN region MAEs were 94.0±10.6 and 102.9±14.7HU for pix2pix and CycleGAN, respectively. For patients with different tissue contrast (water mDixon MR images), the MAEs increased to 122.1±6.3 and 132.8±5.5HU for pix2pix and CycleGAN, respectively. Our results suggest that combining overlapping sCT estimations at each voxel reduced both MAE and ME compared to single-view non-overlapping patch results. Absolute percent mean/max dose errors were 2% or less for the PTV and all clinically relevant structures in our independent testing set, including structures with image artifacts. Quantitative DRR comparison between planning CTs and sCTs showed agreement of bony region positions to <1mm. The dosimetric and MAE based accuracy, along with the similarity between DRRs from sCTs, indicate that pix2pix and CycleGAN are promising methods for MR-only treatment planning for HN cancer. Our methods investigated for overlapping patch-based HU estimations also indicate that combining transformation estimations of overlapping patches is a potential method to reduce generation errors while also providing a tool to potentially estimate the MR to CT aleatoric model transformation uncertainty. However, because of small patient sample sizes, further studies are required.

Inefficiency of orientation averaging: Evidence for hybrid serial/parallel temporal integration.

Intuition suggests that increased viewing time should allow for the accumulation of more visual information, but scant support for this idea has been found in studies of voluntary averaging, where observers are asked to make decisions based on perceived average size. In this paper we examine the dynamics of information accrual in an orientation-averaging task. With orientation (unlike intensive dimensions such as size), it is relatively safe to use an item's physical value as an approximation for its average perceived value. We displayed arrays containing eight iso-eccentric Gabor patterns, and asked six trained psychophysical observers to compare their average orientation with that of probe stimuli that were visible before, during, or only after the presentation of the Gabor array. From the relationship between orientation variance and human performance, we obtained estimates of effective set size, i.e., the number of items that an ideal observer would need to assess in order to estimate average orientation as well as our human observers did. We found that display duration had only a modest influence on effective set size. It rose from an average of ∼2 for 0.1-s displays to an average of ∼3 for 3.3-s displays. These results suggest that the visual computation is neither purely serial nor purely parallel. Computations of this nature can be made with a hybrid process that takes a series of subsamples of a few elements at a time.

Discovering intrinsic properties of human observers' visual search and mathematical observers' scanning.

There is a lack of consensus in measuring observer performance in search tasks. To pursue a consensus, we set our goal to obtain metrics that are practical, meaningful, and predictive. We consider a metric practical if it can be implemented to measure human and computer observers' performance. To be meaningful, we propose to discover intrinsic properties of search observers and formulate the metrics to characterize these properties. If the discovered properties allow verifiable predictions, we consider them predictive. We propose a theory and a conjecture toward two intrinsic properties of search observers: rationality in classification as measured by the location-known-exactly (LKE) receiver operating characteristic (ROC) curve and location uncertainty as measured by the effective set size (M*). These two properties are used to develop search models in both single-response and free-response search tasks. To confirm whether these properties are "intrinsic," we investigate their ability in predicting search performance of both human and scanning channelized Hotelling observers. In particular, for each observer, we designed experiments to measure the LKE-ROC curve and M*, which were then used to predict the same observer's performance in other search tasks. The predictions were then compared to the experimentally measured observer performance. Our results indicate that modeling the search performance using the LKE-ROC curve and M* leads to successful predictions in most cases.

A unified Bayesian observer analysis for set size and cueing effects on perceptual decisions and saccades

Visual search and cueing tasks have been employed extensively in attentional research, with each having a standard effect (visual search: set size effects, cueing: cue validity). Generally these effects have been treated with different (but often similar) attentional theories. The present study aims to consolidate cueing and set size effects within an ideal observer approach. Four observers performed a yes/no contrast discrimination of a gaussian signal in a task combining cueing with visual search. The signal appeared in half the trials, and effective set size (M, 2 to 8) was determined by one primary precue (having 50% validity in signal present trials) and M-1 secondary precues. There were two stimulus durations: 1 second (eye movements allowed), and the first-saccade latency (in the 1 second duration condition) minus 80 milliseconds. Simulations found that an ideal observer for the perceptual yes/no decisions and the first saccadic localization decisions predicted both set size and cueing effects with a single weighting mechanism, providing a unifying account. For the human observer results, a modified ideal observer (with performance matched to human performance) fit the yes/no perceptual decisions well. For the first saccadic decisions, there was evidence of use of the primary cue, but the modified ideal observer was not a good fit, indicating a suboptimal use of the cue. We discuss possible underlying assumptions about the task that might explain the apparent suboptimal nature of saccadic decisions and the overall utility of the ideal observer for cueing and visual search studies in visual attention and saccades.

Depth and Size Information Reduce Effective Set Size for Visual Search in Real-World Scenes

Depth and Size Information Reduce Effective Set Size for Visual Search in Real-World Scenes

Depth and size information reduce effective set size for visual search in real-world scenes

Depth and size information reduce effective set size for visual search in real-world scenes

Salience and target selection in visual search

Why is it easy to find strawberries and difficult to collect gooseberries or green tomatoes? Why don't we see the tree in the forest but do see the single tree in the garden? Various explanations have been given to account for these phenomena. The present paper is concentrating on salience, a property apparently important for search but not frequently discussed in this context. Salience lets targets stand out and thus controls the selection of items that need to be investigated when a certain target is to be searched for. This proposal was tested in two series of experiments. It was found that salience detection and target identification followed different time courses; even typical “pop-out” targets (bright or dark blobs, gaps) were faster located than identified. The importance of salience in visual search was further investigated by manipulating target salience without affecting stimulus properties that are assumed to be relevant for search. When the salience of individual items was increased, the effective set size was reduced and search performance improved. This suggests an interactive and complementary function of salience and attention which is further discussed.

Why are there eccentricity effects in visual search? Visual and attentional hypotheses.

In standard visual search experiments, observers search for a target item among distracting items. The locations of target items are generally random within the display and ignored as a factor in data analysis. Previous work has shown that targets presented near fixation are, in fact, found more efficiently than are targets presented at more peripheral locations. This paper proposes that the primary cause of this "eccentricity effect" (Carrasco, Evert, Chang, & Katz, 1995) is an attentional bias that allocates attention preferentially to central items. The first four experiments dealt with the possibility that visual, and not attentional, factors underlie the eccentricity effect. They showed that the eccentricity effect cannot be accounted for by the peripheral reduction in visual sensitivity, peripheral crowding, or cortical magnification. Experiment 5 tested the attention allocation model and also showed that RT x set size effects can be independent of eccentricity effects. Experiment 6 showed that the effective set size in a search task depends, in part, on the eccentricity of the target because observers search from fixation outward.