Catalyst improves OOD detection by multiplicatively scaling baseline scores using channel-wise statistics from pre-pooling feature maps, reducing average FPR by 22-33% on standard benchmarks.
TurkerGaze: Crowdsourcing Saliency with Webcam based Eye Tracking
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
abstract
Traditional eye tracking requires specialized hardware, which means collecting gaze data from many observers is expensive, tedious and slow. Therefore, existing saliency prediction datasets are order-of-magnitudes smaller than typical datasets for other vision recognition tasks. The small size of these datasets limits the potential for training data intensive algorithms, and causes overfitting in benchmark evaluation. To address this deficiency, this paper introduces a webcam-based gaze tracking system that supports large-scale, crowdsourced eye tracking deployed on Amazon Mechanical Turk (AMTurk). By a combination of careful algorithm and gaming protocol design, our system obtains eye tracking data for saliency prediction comparable to data gathered in a traditional lab setting, with relatively lower cost and less effort on the part of the researchers. Using this tool, we build a saliency dataset for a large number of natural images. We will open-source our tool and provide a web server where researchers can upload their images to get eye tracking results from AMTurk.
representative citing papers
EIHF injects high-frequency evidence early to reshape class-conditional feature geometry and reduce ID/OOD Mahalanobis overlap for geometry-sensitive OOD detection.
GazeCode uses multi-digit recall tasks with anti-peripheral stimulus design to strengthen label validity in unsupervised mobile gaze data collection.
ConjNorm reframes OOD detection score design as optimizing norm p in an exponential family density model via a Bregman divergence theorem, with a tractable Monte Carlo estimator, claiming SOTA gains on CIFAR-100 and ImageNet-1K.
CoOD decomposes inputs into components and applies Component Shift Score plus Compositional Consistency Score to improve detection of both standard and compositional out-of-distribution data.
TTL dynamically learns OOD textual semantics from unlabeled test streams via prompt updates, purification, and a knowledge bank to improve detection performance in pretrained VLMs.
citing papers explorer
-
Catalyst: Out-of-Distribution Detection via Elastic Scaling
Catalyst improves OOD detection by multiplicatively scaling baseline scores using channel-wise statistics from pre-pooling feature maps, reducing average FPR by 22-33% on standard benchmarks.
-
Early High-Frequency Injection for Geometry-Sensitive OOD Detection
EIHF injects high-frequency evidence early to reshape class-conditional feature geometry and reduce ID/OOD Mahalanobis overlap for geometry-sensitive OOD detection.
-
GazeCode: Recall-Based Verification for Higher-Quality In-the-Wild Mobile Gaze Data Collection
GazeCode uses multi-digit recall tasks with anti-peripheral stimulus design to strengthen label validity in unsupervised mobile gaze data collection.
-
ConjNorm: Tractable Density Estimation for Out-of-Distribution Detection
ConjNorm reframes OOD detection score design as optimizing norm p in an exponential family density model via a Bregman divergence theorem, with a tractable Monte Carlo estimator, claiming SOTA gains on CIFAR-100 and ImageNet-1K.
-
Component-Based Out-of-Distribution Detection
CoOD decomposes inputs into components and applies Component Shift Score plus Compositional Consistency Score to improve detection of both standard and compositional out-of-distribution data.
-
TTL: Test-time Textual Learning for OOD Detection with Pretrained Vision-Language Models
TTL dynamically learns OOD textual semantics from unlabeled test streams via prompt updates, purification, and a knowledge bank to improve detection performance in pretrained VLMs.