EDITS: Enhancing Dataset Distillation with Implicit Textual Semantics
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913119322767360 |
|---|---|
| author | Xia, Qianxin Du, Jiawei Lu, Guoming Shu, Zhiyong Wang, Jielei |
| author_facet | Xia, Qianxin Du, Jiawei Lu, Guoming Shu, Zhiyong Wang, Jielei |
| contents | Dataset distillation aims to synthesize a compact dataset from the original large-scale one, enabling highly efficient learning while preserving competitive model performance. However, traditional techniques primarily capture low-level visual features, neglecting the high-level semantic and structural information inherent in images. In this paper, we propose EDITS, a novel framework that exploits the implicit textual semantics within the image data to achieve enhanced distillation. First, external texts generated by a Vision Language Model (VLM) are fused with image features through a Global Semantic Query module, forming the prior clustered buffer. Local Semantic Awareness then selects representative samples from the buffer to construct image and text prototypes, with the latter produced by guiding a Large Language Model (LLM) with meticulously crafted prompt. Ultimately, Dual Prototype Guidance strategy generates the final synthetic dataset through a diffusion model. Extensive experiments confirm the effectiveness of our method.Source code is available in: https://github.com/einsteinxia/EDITS. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_13858 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | EDITS: Enhancing Dataset Distillation with Implicit Textual Semantics Xia, Qianxin Du, Jiawei Lu, Guoming Shu, Zhiyong Wang, Jielei Computer Vision and Pattern Recognition Dataset distillation aims to synthesize a compact dataset from the original large-scale one, enabling highly efficient learning while preserving competitive model performance. However, traditional techniques primarily capture low-level visual features, neglecting the high-level semantic and structural information inherent in images. In this paper, we propose EDITS, a novel framework that exploits the implicit textual semantics within the image data to achieve enhanced distillation. First, external texts generated by a Vision Language Model (VLM) are fused with image features through a Global Semantic Query module, forming the prior clustered buffer. Local Semantic Awareness then selects representative samples from the buffer to construct image and text prototypes, with the latter produced by guiding a Large Language Model (LLM) with meticulously crafted prompt. Ultimately, Dual Prototype Guidance strategy generates the final synthetic dataset through a diffusion model. Extensive experiments confirm the effectiveness of our method.Source code is available in: https://github.com/einsteinxia/EDITS. |
| title | EDITS: Enhancing Dataset Distillation with Implicit Textual Semantics |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.13858 |