TurkerGaze: Crowdsourcing Saliency with Webcam based Eye Tracking
arXiv:1504.06755
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.
9 pages, 14 figures
Cited by in corpus (40)
- Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks
- Energy-based Out-of-distribution Detection
- Learning Confidence for Out-of-Distribution Detection in Neural Networks
- Contextual Encoder-Decoder Network for Visual Saliency Prediction
- Learning Visual Importance for Graphic Designs and Data Visualizations
- Learning to Find Eye Region Landmarks for Remote Gaze Estimation in Unconstrained Settings
- Generative Probabilistic Novelty Detection with Adversarial Autoencoders
- Shallow and Deep Convolutional Networks for Saliency Prediction
- SaltiNet: Scan-path Prediction on 360 Degree Images using Saliency Volumes
- Eye Tracking for Everyone
- Seeing with Humans: Gaze-Assisted Neural Image Captioning
- Detecting Out-of-Distribution Inputs in Deep Neural Networks Using an Early-Layer Output
- Robust Out-of-distribution Detection for Neural Networks
- Open-set Label Noise Can Improve Robustness Against Inherent Label Noise
- Saliency Prediction in the Deep Learning Era: Successes, Limitations, and Future Challenges
- State-of-the-Art in Human Scanpath Prediction
- Learning Uncertain Convolutional Features for Accurate Saliency Detection
- PathGAN: Visual Scanpath Prediction with Generative Adversarial Networks
- Hyperparameter-Free Out-of-Distribution Detection Using Softmax of Scaled Cosine Similarity
- Heatmap-based Out-of-Distribution Detection
- Entity Recognition at First Sight: Improving NER with Eye Movement Information
- Context Proposals for Saliency Detection
- Unified Classification and Rejection: A One-versus-All Framework
- Unsupervised Out-of-Distribution Detection by Maximum Classifier Discrepancy
- ATOM: Robustifying Out-of-distribution Detection Using Outlier Mining
- MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation
- Multiscale Score Matching for Out-of-Distribution Detection
- DICE: Diversity in Deep Ensembles via Conditional Redundancy Adversarial Estimation
- WEPSAM: Weakly Pre-Learnt Saliency Model
- Thinkback: Task-SpecificOut-of-Distribution Detection
- Single Layer Predictive Normalized Maximum Likelihood for Out-of-Distribution Detection
- Exploring Covariate and Concept Shift for Detection and Calibration of Out-of-Distribution Data
- End-to-End Saliency Mapping via Probability Distribution Prediction
- Saliency Revisited: Analysis of Mouse Movements versus Fixations
- Personalized Saliency and its Prediction
- Bridging In- and Out-of-distribution Samples for Their Better Discriminability
- Out-of-Distribution Detection Using an Ensemble of Self Supervised Leave-out Classifiers
- MOOD: Multi-level Out-of-distribution Detection
- FastSal: a Computationally Efficient Network for Visual Saliency Prediction
- Revealing the Distributional Vulnerability of Discriminators by Implicit Generators