# Debiasing Central Fixation Confounds Reveals a Peripheral "Sweet Spot" for Human-like Scanpaths in Hard-Attention Vision

arXiv: https://arxiv.org/abs/2602.14834 · HTML: https://mlsanity.com/paper/2602.14834/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 21 references that are in the map.
- First posted on arXiv 16 Feb 2026.

## Builds on (most cited first, 5 of 5)

- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [Recurrent Models of Visual Attention](https://mlsanity.com/paper/1406.6247/) (2014)
- [FixaTons: A collection of Human Fixations Datasets and Metrics for Scanpath Similarity](https://mlsanity.com/paper/1802.02534/) (2018)
- [Gaze-Guided Learning: Avoiding Shortcut Bias in Visual Classification](https://mlsanity.com/paper/2504.05583/) (2025)
- [Emergence of Fixational and Saccadic Movements in a Multi-Level Recurrent Attention Model for Vision](https://mlsanity.com/paper/2505.13191/) (2025)

## Built on by (most cited first, 1 of 1)

- [EVA: Bridging Performance and Human Alignment in Hard-Attention Vision Models for Image Classification](https://mlsanity.com/paper/2603.27340/) (2026)

## Abstract (arXiv)

Human eye movements in visual recognition reflect a balance between foveal sampling and peripheral context. Task-driven hard-attention models for vision are often evaluated by how well their scanpaths match human gaze. However, common scanpath metrics can be strongly confounded by dataset-specific center bias, especially on object-centric datasets. Using Gaze-CIFAR-10, we show that a trivial center-fixation baseline achieves surprisingly strong scanpath scores, approaching many learned policies. This makes standard metrics optimistic and blurs the distinction between genuine behavioral alignment and mere central tendency. We then analyze a hard-attention classifier under constrained vision by sweeping foveal patch size and peripheral context, revealing a peripheral sweet spot: only a narrow range of sensory constraints yields scanpaths that are simultaneously (i) above the center baseline after debiasing and (ii) temporally human-like in movement statistics. To address center bias, we propose GCS (Gaze Consistency Score), a center-debiased composite metric augmented with movement similarity. GCS uncovers a robust sweet spot at medium patch size with both foveal and peripheral vision, that is not obvious from raw scanpath metrics or accuracy alone, and also highlights a "shortcut regime" when the field-of-view becomes too large. We discuss implications for evaluating active perception on object-centric datasets and for designing gaze benchmarks that better separate behavioral alignment from center bias.