GSE: Evaluating Sticker Visual Semantic Similarity via a General Sticker Encoder

• URL: http://arxiv.org/abs/2511.04977v1
• Date: Fri, 07 Nov 2025 04:29:16 GMT
• Title: GSE: Evaluating Sticker Visual Semantic Similarity via a General Sticker Encoder
• Abstract summary: Stickers have become a popular form of visual communication, yet understanding their semantic relationships remains challenging.<n>We formally define the Sticker Semantic Similarity task and introduce Triple-S, the first benchmark for this task.<n>We show that existing pretrained vision and multimodal models struggle to capture nuanced sticker semantics.<n>We propose the General Sticker (GSE), a lightweight and versatile model that learns robust sticker embeddings using both Triple-S and additional datasets.
• Score: 25.86712652829893
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Stickers have become a popular form of visual communication, yet understanding their semantic relationships remains challenging due to their highly diverse and symbolic content. In this work, we formally {define the Sticker Semantic Similarity task} and introduce {Triple-S}, the first benchmark for this task, consisting of 905 human-annotated positive and negative sticker pairs. Through extensive evaluation, we show that existing pretrained vision and multimodal models struggle to capture nuanced sticker semantics. To address this, we propose the {General Sticker Encoder (GSE)}, a lightweight and versatile model that learns robust sticker embeddings using both Triple-S and additional datasets. GSE achieves superior performance on unseen stickers, and demonstrates strong results on downstream tasks such as emotion classification and sticker-to-sticker retrieval. By releasing both Triple-S and GSE, we provide standardized evaluation tools and robust embeddings, enabling future research in sticker understanding, retrieval, and multimodal content generation. The Triple-S benchmark and GSE have been publicly released and are available here.

Related papers

• Visually Prompted Benchmarks Are Surprisingly Fragile [82.98001690512461]
  Key challenge in evaluating VLMs is testing their ability to analyze visual content independently from their textual priors.<n>We demonstrate how details in benchmark setup, including visual marker design and dataset size, have a significant influence on model performance and leaderboard rankings.<n>To mitigate this instability, we curate existing datasets to create VPBench, a larger visually prompted benchmark with 16 visual marker variants.
  arXiv  Detail & Related papers  (2025-12-19T18:26:58Z)
• A Generative Framework for Personalized Sticker Retrieval [73.57899194210141]
  We propose PEARL, a novel generative framework for personalized sticker retrieval.<n>We make two key contributions: (i) To encode user-specific sticker preferences, we design a representation learning model to learn discriminative user representations, and (ii) To generate stickers aligned with a user's query intent, we propose a novel intent-aware learning objective.<n> Empirical results from both offline evaluations and online tests demonstrate that PEARL significantly outperforms state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-09-22T13:11:44Z)
• Small Stickers, Big Meanings: A Multilingual Sticker Semantic Understanding Dataset with a Gamified Approach [25.86712652829893]
  We introduce Sticktionary, a gamified annotation framework designed to gather diverse, high-quality, and contextually resonant sticker queries.<n>Second, we present StickerQueries, a multilingual sticker query dataset containing 1,115 English and 615 Chinese queries, annotated by over 60 contributors across 60+ hours.<n>Third, we demonstrate that our approach significantly enhances query generation quality, retrieval accuracy, and semantic understanding in the sticker domain.
  arXiv  Detail & Related papers  (2025-06-02T13:38:45Z)
• PerSRV: Personalized Sticker Retrieval with Vision-Language Model [21.279568613306573]
  We propose the Personalized Sticker Retrieval with Vision-Language Model framework, namely PerSRV, structured into offline calculations and online processing modules.<n>For sticker-level semantic understanding, we supervised fine-tuned LLaVA-1.5-7B to generate human-like sticker semantics.<n>Thirdly, we cluster style centroids based on users' historical interactions to achieve personal preference modeling.
  arXiv  Detail & Related papers  (2024-10-29T07:13:47Z)
• HuBERTopic: Enhancing Semantic Representation of HuBERT through Self-supervision Utilizing Topic Model [62.995175485416]
  We propose a new approach to enrich the semantic representation of HuBERT. An auxiliary topic classification task is added to HuBERT by using topic labels as teachers. Experimental results demonstrate that our method achieves comparable or better performance than the baseline in most tasks.
  arXiv  Detail & Related papers  (2023-10-06T02:19:09Z)
• Sticker820K: Empowering Interactive Retrieval with Stickers [34.67442172774095]
  We propose a large-scale Chinese sticker dataset, namely Sticker820K, which consists of 820k image-text pairs. Each sticker has rich and high-quality textual annotations, including descriptions, optical characters, emotional labels, and style classifications. For the text-to-image retrieval task, our StickerCLIP demonstrates strong superiority over the CLIP, which achieves an absolute gain of 66.0% in mean recall.
  arXiv  Detail & Related papers  (2023-06-12T05:06:53Z)
• Selecting Stickers in Open-Domain Dialogue through Multitask Learning [51.67855506570727]
  We propose a multitask learning method comprised of three auxiliary tasks to enhance the understanding of dialogue history, emotion and semantic meaning of stickers. Our model can better combine the multimodal information and achieve significantly higher accuracy over strong baselines.
  arXiv  Detail & Related papers  (2022-09-16T03:45:22Z)
• Learning to Respond with Your Favorite Stickers: A Framework of Unifying Multi-Modality and User Preference in Multi-Turn Dialog [67.91114640314004]
  Stickers with vivid and engaging expressions are becoming increasingly popular in online messaging apps. Some works are dedicated to automatically select sticker response by matching the stickers image with previous utterances. We propose to recommend an appropriate sticker to user based on multi-turn dialog context and sticker using history of user.
  arXiv  Detail & Related papers  (2020-11-05T03:31:17Z)
• Learning to Respond with Stickers: A Framework of Unifying Multi-Modality in Multi-Turn Dialog [65.7021675527543]
  Stickers with vivid and engaging expressions are becoming increasingly popular in online messaging apps. Some works are dedicated to automatically select sticker response by matching text labels of stickers with previous utterances. We propose to recommend an appropriate sticker to user based on multi-turn dialog context history without any external labels.
  arXiv  Detail & Related papers  (2020-03-10T13:10:26Z)

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.