---
title: "S2TDR: Structured Multimodal Reconstruction for Sticker-Aware Dialogue Understanding"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2610.41176v1"
permalink: "https://feed.astralor.com/items/itm_41efc66c9393b60d2d589257"
published: "2026-10-30T02:57:24.000Z"
score: 54
channel: "tech"
tags: ["推理效率","多模态","S2TDR","Structured Multimodal Reconstruction","Sticker-Aware Dialogue Understanding","Stickers","Their","Sticker-to-Text Dialogue Reconstruction"]
---

# S2TDR: Structured Multimodal Reconstruction for Sticker-Aware Dialogue Understanding

## At a glance

Stickers are widely used in social media dialogues to express affect, stance, and intent beyond literal text. Their meanings are often non-literal and context-dependent, making them difficult to handle with generic visu…

## Key points

- Stickers are widely used in social media dialogues to express affect, stance, and intent beyond literal text.
- Their meanings are often non-literal and context-dependent, making them difficult to handle with generic visual fusion or standalone captioning.
- We propose Sticker-to-Text Dialogue Reconstruction (S2TDR), a structured multimodal reconstruction framework that converts sticker turns into contextually appropriate textual utterances.

## Full text

Stickers are widely used in social media dialogues to express affect, stance, and intent beyond literal text. Their meanings are often non-literal and context-dependent, making them difficult to handle with generic visual fusion or standalone captioning. We propose Sticker-to-Text Dialogue Reconstruction (S2TDR), a structured multimodal reconstruction framework that converts sticker turns into contextually appropriate textual utterances. S2TDR decomposes sticker interpretation into visual cue extraction, affective-intent inference, and context-aware dialogue reconstruction, using OCR text only as an auxiliary cue when available. The reconstructed dialogue can then be consumed by existing text-based dialogue models. Experiments on sentiment analysis, intent recognition, and dialogue summarization show that S2TDR generally improves multiple downstream models over direct sticker integration methods. Further analyses on component ablations, OCR-empty stickers, and reconstruction quality indicate that structured reconstruction provides an interpretable and effective way to integrate sticker semantics into dialogue understanding.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_41efc66c9393b60d2d589257