GSE: Evaluating Sticker Visual Semantic Similarity via a General Sticker Encoder
This work addresses the challenge of sticker semantic understanding for applications in visual communication and multimodal AI, providing standardized tools for future research.
The paper tackles the problem of understanding semantic relationships between stickers by defining the Sticker Semantic Similarity task and introducing Triple-S, the first benchmark with 905 annotated sticker pairs, and proposes the General Sticker Encoder (GSE), which achieves superior performance on unseen stickers and strong results on downstream tasks like emotion classification and retrieval.
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.