Research map / Vision-language and multimodal / CLIP and zero-shot
Generalized Decoding for Pixel, Image, and Language
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Pretrain jointly with retrieval, captioning, and referring segmentation (ablation) | X-Decoder · ADE | mIoU: 23.4 -> 21.8 | Helped |
| Pretrain with image captioning (ablation) | X-Decoder · g-Ref | cIoU: 59.7 -> 57.7 | Helped |
| Pretrain with referring segmentation (ablation) | X-Decoder · COCO-Karpathy | CIDEr: 82.0 -> 78.8 | Helped |
| Allow global and fine-grained latent image queries to inform caption text queries (ablation) | X-Decoder · COCO-Karpathy | CIDEr: 82.0 -> 78.6; 82.0 -> 78.9 | Helped |
| Allow text queries to condition latent queries (ablation) | X-Decoder · g-Ref | cIoU: 59.7 -> 57.6 | Helped |
| Add deformable attention (ablation) | X-Decoder (L) · ADE open-vocabulary segmentation | mIoU: 29.6 -> 26.6 | Hurt |
| Add deformable attention (ablation) | X-Decoder (L) · COCO segmentation | PQ: 56.9 -> 57.4 | Helped |
| Remove Visual Genome from VL pretraining data (ablation) | X-Decoder · COCO-Karpathy | CIDEr: 85.0 -> 107.1 | Mixed |
| Pretrain on image-text pairs alongside segmentation (ablation) | X-Decoder · ADE-150 | mIoU: 19.5 -> 25.0 | Helped |
| Use noun phrases extracted from COCO captions as extra supervision (ablation) | X-Decoder-Seg · ADE-150 | mIoU: 19.5 -> 23.8 | Helped |
| Use a VL batch size of 512 instead of 1024 (ablation) | X-Decoder · COCO-Karpathy | CIDEr: 85.0 -> 76.1 | Hurt |
Closest papers by meaning
- SED: A Simple Encoder-Decoder for Open-Vocabulary Semantic Segmentation (CVPR 2024 · similarity 0.90)
- Rethinking Decoders for Transformer-based Semantic Segmentation: A Compression Perspective (NeurIPS 2024 · similarity 0.89)
- Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (CVPR 2024 · similarity 0.89)
- Text4Seg: Reimagining Image Segmentation as Text Generation (ICLR 2025 · similarity 0.88)
- Pixels to Tokens: Token Space Efficient Active Learning for Low-Budget Semantic Segmentation (NeurIPS 2026 · similarity 0.88)
- Test-Time Adaptation of Vision-Language Models for Open-Vocabulary Semantic Segmentation (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.