# FOCUS: Fused Observation of Channels for Unveiling Spectra

arXiv: https://arxiv.org/abs/2507.14787 · HTML: https://mlsanity.com/paper/2507.14787/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 33 of its 50 references that are in the map.
- First posted on arXiv 20 Jul 2025.

## Builds on (most cited first, 30 of 33)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networks](https://mlsanity.com/paper/1710.11063/) (2017)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [RISE: Randomized Input Sampling for Explanation of Black-box Models](https://mlsanity.com/paper/1806.07421/) (2018)
- [Quantifying Attention Flow in Transformers](https://mlsanity.com/paper/2005.00928/) (2020)
- [Going deeper with Image Transformers](https://mlsanity.com/paper/2103.17239/) (2021)
- [This Looks Like That: Deep Learning for Interpretable Image Recognition](https://mlsanity.com/paper/1806.10574/) (2018)
- [Transformer Interpretability Beyond Attention Visualization](https://mlsanity.com/paper/2012.09838/) (2020)
- [Score-CAM: Score-Weighted Visual Explanations for Convolutional Neural Networks](https://mlsanity.com/paper/1910.01279/) (2019)
- [DeepViT: Towards Deeper Vision Transformer](https://mlsanity.com/paper/2103.11886/) (2021)
- [Eigen-CAM: Class Activation Map using Principal Components](https://mlsanity.com/paper/2008.00299/) (2020)
- [Adaptive Token Sampling For Efficient Vision Transformers](https://mlsanity.com/paper/2111.15667/) (2021)
- [SpectralFormer: Rethinking Hyperspectral Image Classification with Transformers](https://mlsanity.com/paper/2107.02988/) (2021)
- [Axiom-based Grad-CAM: Towards Accurate Visualization and Explanation of CNNs](https://mlsanity.com/paper/2008.02312/) (2020)
- [Smooth Grad-CAM++: An Enhanced Inference Level Visualization Technique for Deep Convolutional Neural Network Models](https://mlsanity.com/paper/1908.01224/) (2019)
- [HyperPrompt: Prompt-based Task-Conditioning of Transformers](https://mlsanity.com/paper/2203.00759/) (2022)
- [ProtoPFormer: Concentrating on Prototypical Parts in Vision Transformers for Interpretable Image Recognition](https://mlsanity.com/paper/2208.10431/) (2022)
- [Zero-TPrune: Zero-Shot Token Pruning through Leveraging of the Attention Graph in Pre-Trained Transformers](https://mlsanity.com/paper/2305.17328/) (2023)
- [Transferable Adversarial Attacks on Vision Transformers with Token Gradient Regularization](https://mlsanity.com/paper/2303.15754/) (2023)
- [Optimizing Relevance Maps of Vision Transformers Improves Robustness](https://mlsanity.com/paper/2206.01161/) (2022)
- [Visual Instance-aware Prompt Tuning](https://mlsanity.com/paper/2507.07796/) (2025)
- [HSI-CNN: A Novel Convolution Neural Network for Hyperspectral Image](https://mlsanity.com/paper/1802.10478/) (2018)
- [A Simple Interpretable Transformer for Fine-Grained Image Classification and Analysis](https://mlsanity.com/paper/2311.04157/) (2023)
- [Visual Variational Autoencoder Prompt Tuning](https://mlsanity.com/paper/2503.17650/) (2025)
- [SANet: Structure-Aware Network for Visual Tracking](https://mlsanity.com/paper/1611.06878/) (2016)
- [HSViT: Horizontally Scalable Vision Transformer](https://mlsanity.com/paper/2404.05196/) (2024)
- [SpecFormer: Guarding Vision Transformer Robustness via Maximum Singular Value Penalization](https://mlsanity.com/paper/2402.03317/) (2024)
- [SSA: Semantic Structure Aware Inference for Weakly Pixel-Wise Dense Predictions without Cost](https://mlsanity.com/paper/2111.03392/) (2021)

## Built on by (most cited first, 7 of 7)

- [Prompt-based Adaptation in Large-scale Vision Models: A Survey](https://mlsanity.com/paper/2510.13219/) (2025)
- [Not All Directions Matter: Towards Structured and Task-Aware Low-Rank Model Adaptation](https://mlsanity.com/paper/2603.14228/) (2026)
- [Attention Sink in Transformers: A Survey on Utilization, Interpretation, and Mitigation](https://mlsanity.com/paper/2604.10098/) (2026)
- [CTR-LoRA: Curvature-Aware and Trust-Region Guided Low-Rank Adaptation for Large Language Models](https://mlsanity.com/paper/2510.15962/) (2025)
- [Self-Supervised Visual Prompting for Cross-Domain Road Damage Detection](https://mlsanity.com/paper/2511.12410/) (2025)
- [Layer-Specific Prompt Fusion Discovery via Differentiable Search in Vision Foundation Models](https://mlsanity.com/paper/2606.26379/) (2026)
- [TASAM: Terrain-and-Aware Segment Anything Model for Temporal-Scale Remote Sensing Segmentation](https://mlsanity.com/paper/2509.15795/) (2025)

## Abstract (arXiv)

Hyperspectral imaging (HSI) captures hundreds of narrow, contiguous wavelength bands, making it a powerful tool in biology, agriculture, and environmental monitoring. However, interpreting Vision Transformers (ViTs) in this setting remains largely unexplored due to two key challenges: (1) existing saliency methods struggle to capture meaningful spectral cues, often collapsing attention onto the class token, and (2) full-spectrum ViTs are computationally prohibitive for interpretability, given the high-dimensional nature of HSI data. We present FOCUS, the first framework that enables reliable and efficient spatial-spectral interpretability for frozen ViTs. FOCUS introduces two core components: class-specific spectral prompts that guide attention toward semantically meaningful wavelength groups, and a learnable [SINK] token trained with an attraction loss to absorb noisy or redundant attention. Together, these designs make it possible to generate stable and interpretable 3D saliency maps and spectral importance curves in a single forward pass, without any gradient backpropagation or backbone modification. FOCUS improves band-level IoU by 15 percent, reduces attention collapse by over 40 percent, and produces saliency results that align closely with expert annotations. With less than 1 percent parameter overhead, our method makes high-resolution ViT interpretability practical for real-world hyperspectral applications, bridging a long-standing gap between black-box modeling and trustworthy HSI decision-making.