Selective Aggregation of Attention Maps Improves Diffusion-Based Visual Interpretation

• URL: http://arxiv.org/abs/2604.05906v1
• Date: Tue, 07 Apr 2026 14:09:00 GMT
• Title: Selective Aggregation of Attention Maps Improves Diffusion-Based Visual Interpretation
• Abstract summary: We show that selectively aggregating cross-attention maps from heads most relevant to a target concept can improve visual interpretability.<n>We also find that the most relevant heads capture concept-specific features more accurately than the least relevant ones, and that selective aggregation helps diagnose prompt misinterpretations.
• Score: 12.969353054636272
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Numerous studies on text-to-image (T2I) generative models have utilized cross-attention maps to boost application performance and interpret model behavior. However, the distinct characteristics of attention maps from different attention heads remain relatively underexplored. In this study, we show that selectively aggregating cross-attention maps from heads most relevant to a target concept can improve visual interpretability. Compared to the diffusion-based segmentation method DAAM, our approach achieves higher mean IoU scores. We also find that the most relevant heads capture concept-specific features more accurately than the least relevant ones, and that selective aggregation helps diagnose prompt misinterpretations. These findings suggest that attention head selection offers a promising direction for improving the interpretability and controllability of T2I generation.

Related papers

• Token-level Response-visual Attention Guidance for Multimodal LLMs Knowledge Distillation [36.9736177938003]
  We introduce Token-level Response-visual Attention Guidance (TRAG)<n>TRAG shifts the focus to response-to-vision signals and employs token-specific objectives.<n>Extensive experimental results on multiple benchmarks demonstrate that TRAG significantly outperforms prior distillation baselines.
  arXiv  Detail & Related papers  (2026-07-01T08:03:30Z)
• Making Training-Free Diffusion Segmentors Scale with the Generative Power [118.72472901404814]
  A line of research focuses on adapting a pre-trained diffusion model to semantic segmentation without any further training.<n>We propose two techniques: auto aggregation and per-pixel rescaling, which together enable training-free segmentation to better leverage generative capability.
  arXiv  Detail & Related papers  (2026-03-06T11:35:37Z)
• iSeg: An Iterative Refinement-based Framework for Training-free Segmentation [85.58324416386375]
  We present a deep experimental analysis on iteratively refining cross-attention map with self-attention map. We propose an effective iterative refinement framework for training-free segmentation, named iSeg. Our proposed iSeg achieves an absolute gain of 3.8% in terms of mIoU compared to the best existing training-free approach in literature.
  arXiv  Detail & Related papers  (2024-09-05T03:07:26Z)
• I2AM: Interpreting Image-to-Image Latent Diffusion Models via Bi-Attribution Maps [8.195126516665914]
  Cross-attention mechanisms in image-to-image (I2I) models are underexplored.<n>This paper introduces Image-to-Image Attribution Maps (I2AM), a method that enhances the interpretability of I2I models.<n>I2AM aggregates cross-attention scores across time steps, attention heads, and layers, offering insights into how critical features are transferred between images.
  arXiv  Detail & Related papers  (2024-07-17T06:15:05Z)
• Diffusion Model with Cross Attention as an Inductive Bias for Disentanglement [58.9768112704998]
  Disentangled representation learning strives to extract the intrinsic factors within observed data. We introduce a new perspective and framework, demonstrating that diffusion models with cross-attention can serve as a powerful inductive bias. This is the first work to reveal the potent disentanglement capability of diffusion models with cross-attention, requiring no complex designs.
  arXiv  Detail & Related papers  (2024-02-15T05:07:54Z)
• DiffAugment: Diffusion based Long-Tailed Visual Relationship Recognition [43.01467525231004]
  We introduce DiffAugment -- a method which augments the tail classes in the linguistic space by making use of WordNet. We demonstrate the effectiveness of hardness-aware diffusion in generating visual embeddings for the tail classes. We also propose a novel subject and object based seeding strategy for diffusion sampling which improves the discriminative capability of the generated visual embeddings.
  arXiv  Detail & Related papers  (2024-01-01T21:20:43Z)
• Harnessing Diffusion Models for Visual Perception with Meta Prompts [68.78938846041767]
  We propose a simple yet effective scheme to harness a diffusion model for visual perception tasks. We introduce learnable embeddings (meta prompts) to the pre-trained diffusion models to extract proper features for perception. Our approach achieves new performance records in depth estimation tasks on NYU depth V2 and KITTI, and in semantic segmentation task on CityScapes.
  arXiv  Detail & Related papers  (2023-12-22T14:40:55Z)
• Visual Commonsense based Heterogeneous Graph Contrastive Learning [79.22206720896664]
  We propose a heterogeneous graph contrastive learning method to better finish the visual reasoning task. Our method is designed as a plug-and-play way, so that it can be quickly and easily combined with a wide range of representative methods.
  arXiv  Detail & Related papers  (2023-11-11T12:01:18Z)
• Your "Attention" Deserves Attention: A Self-Diversified Multi-Channel Attention for Facial Action Analysis [12.544285462327839]
  We propose a compact model to enhance the representational and focusing power of neural attention maps. The proposed method is evaluated on two benchmark databases (BP4D and DISFA) for AU detection and four databases (CK+, MMI, BU-3DFE, and BP4D+) for facial expression recognition. It achieves superior performance compared to the state-of-the-art methods.
  arXiv  Detail & Related papers  (2022-03-23T17:29:51Z)
• Improve the Interpretability of Attention: A Fast, Accurate, and Interpretable High-Resolution Attention Model [6.906621279967867]
  We propose a novel Bilinear Representative Non-Parametric Attention (BR-NPA) strategy that captures the task-relevant human-interpretable information. The proposed model can be easily adapted in a wide variety of modern deep models, where classification is involved. It is also more accurate, faster, and with a smaller memory footprint than usual neural attention modules.
  arXiv  Detail & Related papers  (2021-06-04T15:57:37Z)
• Variational Structured Attention Networks for Deep Visual Representation Learning [49.80498066480928]
  We propose a unified deep framework to jointly learn both spatial attention maps and channel attention in a principled manner. Specifically, we integrate the estimation and the interaction of the attentions within a probabilistic representation learning framework. We implement the inference rules within the neural network, thus allowing for end-to-end learning of the probabilistic and the CNN front-end parameters.
  arXiv  Detail & Related papers  (2021-03-05T07:37:24Z)

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.