Modeling Saliency Dataset Bias

• URL: http://arxiv.org/abs/2505.10169v1
• Date: Thu, 15 May 2025 10:55:47 GMT
• Title: Modeling Saliency Dataset Bias
• Authors: Matthias Kümmerer, Harneet Khanuja, Matthias Bethge,
• Abstract summary: Recent advances in image-based saliency prediction are approaching gold standard performance levels on existing benchmarks.<n>We show that predicting fixations across multiple saliency datasets remains challenging due to dataset bias.<n>We propose a novel architecture extending a mostly dataset-agnostic encoder-decoder structure with fewer than 20 dataset-specific parameters.
• Score: 10.364146597632365
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent advances in image-based saliency prediction are approaching gold standard performance levels on existing benchmarks. Despite this success, we show that predicting fixations across multiple saliency datasets remains challenging due to dataset bias. We find a significant performance drop (around 40%) when models trained on one dataset are applied to another. Surprisingly, increasing dataset diversity does not resolve this inter-dataset gap, with close to 60% attributed to dataset-specific biases. To address this remaining generalization gap, we propose a novel architecture extending a mostly dataset-agnostic encoder-decoder structure with fewer than 20 dataset-specific parameters that govern interpretable mechanisms such as multi-scale structure, center bias, and fixation spread. Adapting only these parameters to new data accounts for more than 75% of the generalization gap, with a large fraction of the improvement achieved with as few as 50 samples. Our model sets a new state-of-the-art on all three datasets of the MIT/Tuebingen Saliency Benchmark (MIT300, CAT2000, and COCO-Freeview), even when purely generalizing from unrelated datasets, but with a substantial boost when adapting to the respective training datasets. The model also provides valuable insights into spatial saliency properties, revealing complex multi-scale effects that combine both absolute and relative sizes.

Related papers

• OmniTraj: Pre-Training on Heterogeneous Data for Adaptive and Zero-Shot Human Trajectory Prediction [62.385417528148224]
  We present OmniTraj, a Transformer-based model pre-trained on a large-scale, heterogeneous dataset.<n>Experiments show that explicitly conditioning on the frame rate enables OmniTraj to achieve state-of-the-art zero-shot transfer performance.
  arXiv  Detail & Related papers  (2025-07-31T15:37:09Z)
• Towards Data-Efficient Pretraining for Atomic Property Prediction [51.660835328611626]
  We show that pretraining on a task-relevant dataset can match or surpass large-scale pretraining.<n>We introduce the Chemical Similarity Index (CSI), a novel metric inspired by computer vision's Fr'echet Inception Distance.
  arXiv  Detail & Related papers  (2025-02-16T11:46:23Z)
• A CLIP-Powered Framework for Robust and Generalizable Data Selection [51.46695086779598]
  Real-world datasets often contain redundant and noisy data, imposing a negative impact on training efficiency and model performance. Data selection has shown promise in identifying the most representative samples from the entire dataset. We propose a novel CLIP-powered data selection framework that leverages multimodal information for more robust and generalizable sample selection.
  arXiv  Detail & Related papers  (2024-10-15T03:00:58Z)
• Pedestrian Attribute Recognition: A New Benchmark Dataset and A Large Language Model Augmented Framework [15.991114464911844]
  In the past five years, no large-scale dataset has been opened to the public. This paper proposes a new large-scale, cross-domain pedestrian attribute recognition dataset, MSP60K. It consists of 60,122 images and 57 attribute annotations across eight scenarios.
  arXiv  Detail & Related papers  (2024-08-19T06:19:31Z)
• Under the Hood of Tabular Data Generation Models: Benchmarks with Extensive Tuning [2.5168710814072894]
  We investigate the impact of dataset-specific tuning on five recent model families for tabular data generation through an extensive benchmark on 16 datasets.<n>Our benchmark demonstrates that, for most models, large-scale dataset-specific tuning substantially improves performance compared to the original configurations.
  arXiv  Detail & Related papers  (2024-06-18T07:27:38Z)
• UniTraj: A Unified Framework for Scalable Vehicle Trajectory Prediction [93.77809355002591]
  We introduce UniTraj, a comprehensive framework that unifies various datasets, models, and evaluation criteria. We conduct extensive experiments and find that model performance significantly drops when transferred to other datasets. We provide insights into dataset characteristics to explain these findings.
  arXiv  Detail & Related papers  (2024-03-22T10:36:50Z)
• A Case for Dataset Specific Profiling [0.9023847175654603]
  Data-driven science is an emerging paradigm where scientific discoveries depend on the execution of computational AI models against rich, discipline-specific datasets. With modern machine learning frameworks, anyone can develop and execute computational models that reveal concepts hidden in the data that could enable scientific applications. For important and widely used datasets, computing the performance of every computational model that can run against a dataset is cost prohibitive in terms of cloud resources.
  arXiv  Detail & Related papers  (2022-08-01T18:38:05Z)
• THE Benchmark: Transferable Representation Learning for Monocular Height Estimation [25.872962101146115]
  We propose a new benchmark dataset to study the transferability of height estimation models in a cross-dataset setting. This benchmark dataset includes a newly proposed large-scale synthetic dataset, a newly collected real-world dataset, and four existing datasets from different cities. In this paper, we propose a scale-deformable convolution module to enhance the window-based Transformer for handling the scale-variation problem in the height estimation task.
  arXiv  Detail & Related papers  (2021-12-30T09:40:26Z)
• MSeg: A Composite Dataset for Multi-domain Semantic Segmentation [100.17755160696939]
  We present MSeg, a composite dataset that unifies semantic segmentation datasets from different domains. We reconcile the generalization and bring the pixel-level annotations into alignment by relabeling more than 220,000 object masks in more than 80,000 images. A model trained on MSeg ranks first on the WildDash-v1 leaderboard for robust semantic segmentation, with no exposure to WildDash data during training.
  arXiv  Detail & Related papers  (2021-12-27T16:16:35Z)
• Comparing Test Sets with Item Response Theory [53.755064720563]
  We evaluate 29 datasets using predictions from 18 pretrained Transformer models on individual test examples. We find that Quoref, HellaSwag, and MC-TACO are best suited for distinguishing among state-of-the-art models. We also observe span selection task format, which is used for QA datasets like QAMR or SQuAD2.0, is effective in differentiating between strong and weak models.
  arXiv  Detail & Related papers  (2021-06-01T22:33:53Z)
• Cross-Dataset Collaborative Learning for Semantic Segmentation [17.55660581677053]
  We present a simple, flexible, and general method for semantic segmentation, termed Cross-Dataset Collaborative Learning (CDCL) Given multiple labeled datasets, we aim to improve the generalization and discrimination of feature representations on each dataset. We conduct extensive evaluations on four diverse datasets, i.e., Cityscapes, BDD100K, CamVid, and COCO Stuff, with single-dataset and cross-dataset settings.
  arXiv  Detail & Related papers  (2021-03-21T09:59:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.