TraceCLIP: Recovering Local Semantics from Patch-to-CLS Contributions

• URL: http://arxiv.org/abs/2607.26107v1
• Date: Tue, 28 Jul 2026 11:25:27 GMT
• Title: TraceCLIP: Recovering Local Semantics from Patch-to-CLS Contributions
• Abstract summary: We introduce TraceCLIP, a training-free framework that recovers latent patch-level semantic evidence.<n>We show that TraceCLIP achieves gains of 1.3 to 4.5 points in average mIoU over the strongest prior training-free methods.
• Score: 20.430455131220544
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Dense vision-language understanding, including object localization, region recognition, and open-vocabulary semantic segmentation, requires associating language concepts with spatially grounded visual regions. CLIP provides a strong foundation for these tasks by learning a shared image-text embedding space from large-scale contrastive pre-training. However, its image-level objective aligns text with a CLS-derived global representation, leaving local vision-language correspondence only indirectly constrained. Existing methods either introduce additional supervision, external models, or task-specific adaptation, while training-free approaches mainly recover dense responses from existing patch features without examining where local semantics become most accessible within CLIP. We introduce TraceCLIP, a training-free framework that recovers latent patch-level semantic evidence by isolating the patch-specific terms written into the CLS attention output. TraceCLIP further converts contribution-derived semantic responses into a semantic-geodesic topology gate that calibrates final-layer patch affinity for dense feature reconstruction. Diagnostic experiments show that these contribution features exhibit strong local semantic discrimination and text-conditioned spatial alignment. On eight zero-shot semantic segmentation benchmarks, TraceCLIP achieves gains of 1.3 to 4.5 points in average mIoU over the strongest prior training-free methods across both backbones and background settings, without additional training, external vision foundation models, or region-level supervision. More broadly, these findings suggest that spatially localized semantics may remain accessible within the internal construction of globally aligned representations.

Related papers

• Fine-grained CLIP fine-tuning with self-annotated region alignment [57.28808560528696]
  Contrastive Language-Image Pre-training has been shown to have limitations in its fine-grained dense feature representation.<n>We propose SFF-CLIP (Self-annotated Fine-grained Fine-tuning for CLIP) to boost the fine-grained representation ability in the CLIP fine-tuning.
  arXiv  Detail & Related papers  (2026-07-15T10:06:39Z)
• PCA-Seg: Revisiting Cost Aggregation for Open-Vocabulary Semantic and Part Segmentation [58.1914505657064]
  This paper proposes a simple yet effective parallel cost aggregation (PCA-Seg) paradigm to alleviate the challenge of knowledge interference between class-level semantics and spatial context.<n>Experiments on eight benchmarks show that each parallel block in PCA-Seg adds merely 0.35M parameters while achieving state-of-the-art OSPS performance.
  arXiv  Detail & Related papers  (2026-03-18T09:26:43Z)
• Structure-Aware Feature Rectification with Region Adjacency Graphs for Training-Free Open-Vocabulary Semantic Segmentation [22.409969687852506]
  We propose a structure-aware feature rectification approach that incorporates instance-specific priors derived directly from the image.<n>Our method effectively suppresses segmentation noise, improves region-level consistency, and achieves strong performance on multiple open-vocabulary segmentation benchmarks.
  arXiv  Detail & Related papers  (2025-12-08T10:00:36Z)
• HarmoCLIP: Harmonizing Global and Regional Representations in Contrastive Vision-Language Models [63.87966115136411]
  HarmoCLIP is a novel framework designed to harmonize global and region representations within Contrastive Language-Image Pre-training.<n>To strengthen the representation capability at the local level, our method introduces a novel Region-Language Alignment supervision strategy.
  arXiv  Detail & Related papers  (2025-11-27T16:24:53Z)
• Generalized Decoupled Learning for Enhancing Open-Vocabulary Dense Perception [71.26728044621458]
  DeCLIP is a novel framework that enhances CLIP by decoupling the self-attention module to obtain content'' and context'' features respectively.<n>It consistently achieves state-of-the-art performance across a broad spectrum of tasks, including 2D detection and segmentation, 3D instance segmentation, video instance segmentation, and 6D object pose estimation.
  arXiv  Detail & Related papers  (2025-08-15T06:43:51Z)
• DeCLIP: Decoupled Learning for Open-Vocabulary Dense Perception [21.87721909270275]
  DeCLIP is a novel framework that enhances CLIP with content'' and context'' features respectively.<n>It significantly outperforms existing methods across multiple open-vocabulary dense prediction tasks.
  arXiv  Detail & Related papers  (2025-05-07T13:46:34Z)
• Refining CLIP's Spatial Awareness: A Visual-Centric Perspective [10.936397225984107]
  Contrastive Language-Image Pre-training excels in global alignment with language but exhibits limited sensitivity to spatial information.<n>Recent approaches have introduced Region-Language Alignment to enhance CLIP's performance in dense multimodal tasks.<n>We propose the Spatial Correlation Distillation (SCD) framework, which preserves CLIP's inherent spatial structure and mitigates the above degradation.
  arXiv  Detail & Related papers  (2025-04-03T07:04:56Z)
• Explore the Potential of CLIP for Training-Free Open Vocabulary Semantic Segmentation [38.16802763051431]
  We propose CLIPtrase, a training-free semantic segmentation strategy. It enhances local feature awareness through recalibrated self-correlation among patches. Experiments show that we are 22.3% ahead of CLIP on average on 9 segmentation benchmarks.
  arXiv  Detail & Related papers  (2024-07-11T08:12:16Z)
• Open-Vocabulary Segmentation with Semantic-Assisted Calibration [68.41025728960176]
  We study open-vocabulary segmentation (OVS) through calibrating in-vocabulary and domain-biased embedding space with contextual prior of CLIP.<n>We present a Semantic-assisted CAlibration Network (SCAN) to achieve state-of-the-art performance on open-vocabulary segmentation benchmarks.
  arXiv  Detail & Related papers  (2023-12-07T07:00:09Z)
• Rewrite Caption Semantics: Bridging Semantic Gaps for Language-Supervised Semantic Segmentation [100.81837601210597]
  We propose Concept Curation (CoCu) to bridge the gap between visual and textual semantics in pre-training data. CoCu achieves superb zero-shot transfer performance and greatly boosts language-supervised segmentation baseline by a large margin.
  arXiv  Detail & Related papers  (2023-09-24T00:05:39Z)
• Context-self contrastive pretraining for crop type semantic segmentation [39.81074867563505]
  The proposed Context-Self Contrastive Loss (CSCL) learns an embedding space that makes semantic boundaries pop-up. For crop type semantic segmentation from Satellite Image Time Series (SITS) we find performance at parcel boundaries to be a critical bottleneck. We present a process for semantic segmentation at super-resolution for obtaining crop classes at a more granular level.
  arXiv  Detail & Related papers  (2021-04-09T11:29:44Z)

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.