# Alfa: Attentive Low-Rank Filter Adaptation for Structure-Aware Cross-Domain Personalized Gaze Estimation

arXiv: https://arxiv.org/abs/2603.08445 · HTML: https://mlsanity.com/paper/2603.08445/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 30 references that are in the map.
- First posted on arXiv 9 Mar 2026.

## Builds on (most cited first, 16 of 16)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [Gaze360: Physically Unconstrained Gaze Estimation in the Wild](https://mlsanity.com/paper/1910.10088/) (2019)
- [ETH-XGaze: A Large Scale Dataset for Gaze Estimation under Extreme Head Pose and Gaze Variation](https://mlsanity.com/paper/2007.15837/) (2020)
- [MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation](https://mlsanity.com/paper/1711.09017/) (2017)
- [MiLoRA: Harnessing Minor Singular Components for Parameter-Efficient LLM Finetuning](https://mlsanity.com/paper/2406.09044/) (2024)
- [Few-Shot Adaptive Gaze Estimation](https://mlsanity.com/paper/1905.01941/) (2019)
- [Improving Few-Shot User-Specific Gaze Adaptation via Gaze Redirection Synthesis](https://mlsanity.com/paper/1904.10638/) (2019)
- [Generalizing Gaze Estimation with Outlier-guided Collaborative Adaptation](https://mlsanity.com/paper/2107.13780/) (2021)
- [GazePointAR: A Context-Aware Multimodal Voice Assistant for Pronoun Disambiguation in Wearable Augmented Reality](https://mlsanity.com/paper/2404.08213/) (2024)
- [Domain Adaptation Gaze Estimation by Embedding with Prediction Consistency](https://mlsanity.com/paper/2011.07526/) (2020)
- [Spectral Adapter: Fine-Tuning in Spectral Space](https://mlsanity.com/paper/2405.13952/) (2024)
- [Offset Calibration for Appearance-Based Gaze Estimation via Gaze Decomposition](https://mlsanity.com/paper/1905.04451/) (2019)
- [Test-Time Personalization with Meta Prompt for Gaze Estimation](https://mlsanity.com/paper/2401.01577/) (2024)
- [GazeGen: Gaze-Driven User Interaction for Visual Content Generation](https://mlsanity.com/paper/2411.04335/) (2024)

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

- [Adaptive Physical-Facial Representation Fusion via Subject-Invariant Cross-Modal Prompt Tuning for Video-Based Emotion Recognition](https://mlsanity.com/paper/2605.05694/) (2026)

## Abstract (arXiv)

Pre-trained gaze models learn to identify useful patterns commonly found across users, but subtle user-specific variations (i.e., eyelid shape or facial structure) can degrade model performance. Test-time personalization (TTP) adapts pre-trained models to these user-specific domain shifts using only a few unlabeled samples. Efficient fine-tuning is critical in performing this domain adaptation: data and computation resources can be limited-especially for on-device customization. While popular parameter-efficient fine-tuning (PEFT) methods address adaptation costs by updating only a small set of weights, they may not be taking full advantage of structures encoded in pre-trained filters. To more effectively leverage existing structures learned during pre-training, we reframe personalization as a process to reweight existing features rather than learning entirely new ones. We present Attentive Low-Rank Filter Adaptation (Alfa) to adapt gaze models by reweighting semantic patterns in pre-trained filters. With Alfa, singular value decomposition (SVD) extracts dominant spatial components that capture eye and facial characteristics across users. Via an attention mechanism, we need only a few unlabeled samples to adjust and reweight pre-trained structures, selectively amplifying those relevant to a target user. Alfa achieves the lowest average gaze errors across four cross-dataset gaze benchmarks, outperforming existing TTP methods and low-rank adaptation (LoRA)-based variants. We also show that Alfa's attentive low-rank methods can be applied to applications beyond vision, such as diffusion-based language models.