华中科技大学白翔团队联合华南理工大学、阿德莱德大学和字节跳动联合推出新一代OCR评测基准OCRBench v2,并发布最新私有数据榜单(2025年9月)。
Seed1.6-vision、Qwen3-Omni-30B-A3B-Instruct和Gemini-2.5-Pro包揽了中英榜单前三名,即便是这些最先进的模型,其平均分也仅勉强达到“60分”的及格线,它们仍难以完全满足复杂多样的真实应用需求。
OCRBench v2榜单测试范围广泛,囊括了从2023年至2025年间的全球58个主流LMMs,有来自谷歌(Gemini 2.5 Pro)、字节跳动(Seed1.6-vision)、OpenAI(GPT-5)的闭源模型,也有来自阿里巴巴(Qwen-VL)、上海人工智能实验室(InternVL)的优秀开源LMMs,能看到LMMs在OCR任务上的表现有着显著进步。
榜单给出了LMMs在文本定位、知识推理等8个核心能力维度的细分指标,直观量化了模型在各类应用场景上的真实表现。
| Rank | Method | Venue | Open-source | LLM Size | Average | Recognition | Referring | Spotting | Extraction | Parsing | Calculation | Understanding | Reasoning |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Seed1.6-vision | No | 62.2 | 70.1 | 59.8 | 38.0 | 89.0 | 22.3 | 84 | 76.4 | 58.0 | ||
| 2 | Qwen3-Omni-30B-A3B Instruct | Arxiy 2025 | Yes | 30B | 61.3 | 72.3 | 62.0 | 45.6 | 93.5 | 20.8 | 67.0 | 74.1 | 55.3 |
| 3 | Gemini-2.5-Pro | No | 59.3 | 70.9 | 45.8 | 13.4 | 93.7 | 26.9 | 84.6 | 75.8 | 63.0 | ||
| 4 | Liama-3.1-Memotron-Nano VL-8B-V1 | Yes | 8B | 56.4 | 62.9 | 61.3 | 68.6 | 88.2 | 10.0 | 44.1 | 75.3 | 41.0 | |
| 5 | GPT5-2025-08-07 | No | 55.5 | 69.3 | 28.2 | 6.5 | 90.1 | 24.8 | 83.7 | 76.4 | 65.4 | ||
| 6 | Ovis2.5-8B | Arxiv 2025 | Yes | 8B | 54.1 | 63.6 | 31.2 | 1.7 | 89.8 | 24.8 | 85.4 | 75.3 | 61.0 |
| 7 | Gemini1.5-Pro | Arxiv 2024 | No | 51.6 | 59.1 | 41.2 | 6.6 | 89.5 | 22.4 | 54.7 | 78.8 | 60.3 | |
| 8 | SAIL-VL2-8B | Arxiv 2025 | Yes | 8B | 49.3 | 67.4 | 30.4 | 2.8 | 91.0 | 23.7 | 51.6 | 75.6 | 52.0 |
| 9 | MiniCPM-V-4.5-8B | Arxiv 2025 | Yes | 8B | 48.4 | 60.3 | 25.7 | 2.2 | 91.1 | 23.4 | 54.1 | 78.6 | 52.0 |
| 10 | GPT-4o | Arxiv 2024 | No | 47.6 | 58.6 | 23.4 | 0.0 | 87.4 | 23.1 | 51.6 | 74.4 | 62.3 | |
| 11 | Claude3.5-sonnet | No | 47.5 | 52.9 | 24.9 | 2.5 | 86.9 | 23.8 | 61.4 | 74.4 | 53.0 | ||
| 12 | InternVL3.5-14B | Arxiv 2025 | Yes | 14B | 47.1 | 50.4 | 24.2 | 1.7 | 88.0 | 23.2 | 61.0 | 72.6 | 55.8 |
| 13 | Step-1V | No | 46.8 | 56.7 | 27.4 | 2.6 | 86.3 | 33.3 | 42.6 | 76.6 | 48.7 | ||
| 14 | InterVL3-14B | Yes | 14B | 46.8 | 55.8 | 24.5 | 2.1 | 89.3 | 21.0 | 59.5 | 72.0 | 50.0 | |
| 15 | Ovis2-8B | Yes | 7B | 46.1 | 54.2 | 20.9 | 0.0 | 83.6 | 24.2 | 54.7 | 74.1 | 57.3 | |
| 16 | InterVL3.5-8B | Arxiv 2025 | Yes | 8B | 46.0 | 49.6 | 25.1 | 0.8 | 86.0 | 21.5 | 63.9 | 75.5 | 45.5 |
| 17 | LLavA-OneVision-1.5-8B-Instruct | Yes | 8B | 46.0 | 54.9 | 22.6 | 0 | 84.4 | 23.5 | 58.0 | 74.5 | 50.3 | |
| 18 | IntemVL3-8B | Yes | 8B | 45.3 | 49.7 | 22.3 | 0.2 | 86.8 | 22.4 | 57.0 | 70.7 | 53.0 | |
| 19 | Grok4 | No | 45.0 | 49.7 | 20.4 | 3.6 | 63.0 | 22.4 | 77.8 | 71.0 | 52.0 | ||
| 20 | GPT-4o-mini | No | 44.1 | 55.3 | 21.8 | 0.0 | 85.4 | 20.6 | 45.2 | 75.5 | 49.0 | ||
| 21 | SAIL-VL-1.6-8B | Arxiv 2025 | Yes | 8B | 43.1 | 56.7 | 24.1 | 2.2 | 79.3 | 22.8 | 45.4 | 69.2 | 45.3 |
| 22 | InterVL2.5-26B | Anxiv 2024 | Yes | 20B | 42.6 | 53.5 | 21.4 | 0 | 84.0 | 21.4 | 51.5 | 67.5 | 41.5 |
| 23 | WeThink-Qwen2.5VL-7B | Anxiv 2025 | Yes | 7B | 42.5 | 59.0 | 23.7 | 3.6 | 34.0 | 25.2 | 71.9 | 73.4 | 49.0 |
| 24 | Claude-sonnet.4-2025051 | No | 42.4 | 56.2 | 24.6 | 1.4 | 78.1 | 24.7 | 36.8 | 78.1 | 39.5 | ||
| 25 | Qwen2-V-7B | Arxiy 2024 | Yes | 8B | 42.3 | 47.0 | 42.0 | 1.5 | 90.2 | 13.7 | 36.4 | 71.1 | 36.6 |
| 26 | Qwen2.5-VL-7B | Arxiv 2025 | Yes | 8B | 41.8 | 51.5 | 24.5 | 3.1 | 64.8 | 13.1 | 53.3 | 78.6 | 45.5 |
| 27 | InterVL2-26B | SCIS 2024 | Yes | 20B | 41.8 | 56.0 | 21.2 | 0 | 80.5 | 23.9 | 40.3 | 72.1 | 40.7 |
| 28 | MiniCPM-o-2.6 | Yes | 8B | 41.6 | 54.1 | 24.7 | 0.3 | 74.4 | 17.6 | 39.2 | 75.7 | 47.0 | |
| 29 | Deepseek-VL2-Small | Arxiv 2024 | Yes | 16B | 41.0 | 56.6 | 23.7 | 0 | 86.4 | 18.9 | 30.6 | 72.2 | 39.5 |
| 30 | InternVL2.5-8B | Arxiy 2024 | Yes | 8B | 40.5 | 48.9 | 21.2 | 0 | 82.1 | 20.3 | 41.2 | 67.8 | 42.3 |
| 31 | Pixtral-12B | Arxiv 2024 | Yes | 12B | 38.4 | 45.1 | 21.8 | 0 | 71.6 | 21.7 | 30.4 | 77.3 | 39.5 |
| 32 | Phi-4-MultiModal | Arxiv 2025 | Yes | 5.6B | 38.1 | 58.4 | 19.0 | 0 | 53.5 | 38.7 | 28.7 | 66.8 | 39.8 |
| 33 | Ovis1.6-3B | Arxiv 2024 | Yes | 3B | 38.0 | 48.5 | 19.5 | 0 | 69.2 | 20.7 | 22.1 | 74.6 | 49.5 |
| 34 | GLM-4v-9B | Arxiv 2024 | Yes | 9B | 37.1 | 52.7 | 20.6 | 0 | 79.4 | 15.9 | 21.5 | 74.7 | 32.0 |
| 35 | IntermVL2-8B | SCIS 2024 | Yes | 8B | 36.1 | 43.0 | 21.6 | 0 | 70.2 | 19.2 | 35.6 | 65.9 | 33.6 |
| 36 | Molmo-7B | CVPR2025 | Yes | 8B | 33.9 | 40.8 | 19.5 | 0 | 51.7 | 10.0 | 33.9 | 67.0 | 48.0 |
| 37 | XComposer2-4KHD | NIPS 2025 | Yes | 7B | 33.9 | 39.5 | 12.0 | 0 | 69.7 | 26.0 | 20.2 | 68.2 | 35.8 |
| 38 | LLaVA-OV-7B | Arxiv 2024 | Yes | 8B | 33.7 | 45.4 | 18.5 | 0 | 60.0 | 15.5 | 32.0 | 59.0 | 39.3 |
| 39 | MiniCPM-V-2.6 | Arxiv 2024 | Yes | 8B | 33.0 | 52.2 | 18.6 | 0.3 | 45.8 | 19.6 | 20.9 | 68.9 | 37.3 |
| 40 | Cambrian-1-8B | NIPS 2025 | Yes | 8B | 32.3 | 44.0 | 19.0 | 0 | 52.3 | 19.0 | 20.7 | 64.0 | 39.3 |
| 41 | Kimi-VL-A3B-16B | Arxiy 2025 | Yes | 16B | 32.1 | 49.1 | 13.5 | 0 | 28.8 | 21.9 | 37.6 | 69.4 | 36.2 |
| 42 | LLaVA-Next-8B | Yes | 8B | 28.5 | 41.4 | 17.0 | 0 | 49.0 | 12.9 | 16.1 | 60.9 | 30.5 | |
| 43 | ldefics3-8B | NeurIPS 2024 Workshop | Yes | 8B | 26.0 | 37.4 | 13.0 | 0 | 28.9 | 19.4 | 21.1 | 65.4 | 21.8 |
| 44 | Eagle-X5-7B | ICLR 2025 | Yes | 8B | 25.7 | 34.6 | 18.5 | 9.7 | 18.5 | 24.0 | 63.1 | 37.0 | |
| 45 | Qwen-VL-chat | Arxiv 2023 | Yes | 8B | 25.7 | 34.1 | 12.6 | 0.1 | 42.6 | 19.5 | 18.4 | 58.3 | 20.3 |
| 46 | Qwen-VL | Arxiv 2023 | Yes | 8B | 24.8 | 35.9 | 4.2 | 0 | 38.7 | 28.5 | 13.8 | 60.1 | 16.9 |
| 47 | Deepseek.VL-7B | Arxiv 2024 | Yes | 7B | 24.5 | 33.5 | 13.7 | 0 | 19.1 | 11.7 | 24.8 | 60.5 | 32.5 |
| 48 | Monkey | CVPR 2024 | Yes | 8B | 24.2 | 31.5 | 0.1 | 0 | 34.4 | 26.3 | 17.7 | 61.4 | 22.4 |
| 49 | DocOwt2 | Arxiv 2024 | Yes | 7B | 23.4 | 25.4 | 7.5 | 0 | 47.1 | 26.2 | 8.3 | 52.8 | 19.5 |
| 50 | TextMonkey | Arxiv 2024 | Yes | 8B | 23.4 | 39.8 | 1.6 | 0 | 27.6 | 24.8 | 10.2 | 62.3 | 21.2 |
| 51 | VILA1.5-8B | CVPR2024 | Yes | 8B | 23.2 | 36.0 | 14.5 | 0 | 26.0 | 17.4 | 20.3 | 44.7 | 27.0 |
| 52 | EMU2-chat | CVPR 2024 | Yes | 37B | 20.2 | 34.3 | 0 | 0 | 20.4 | 21.3 | 20.3 | 47.1 | 18.3 |
| 53 | CogVLM-chat | NIPS 2024 | Yes | 7B | 19.9 | 40.8 | 0 | 0 | 1.6 | 18.6 | 10.9 | 60.2 | 26.8 |
| 54 | YI-VL-6B | Arxiy 2024 | Yes | 6B | 19.7 | 31.1 | 4.0 | 0 | 23.4 | 22.5 | 18.1 | 43.0 | 15.5 |
| 55 | mPLUG-Owl3 | Arxiv 2024 | Yes | 8B | 16.5 | 34.9 | 17.0 | 0 | 12.0 | 14.9 | 24.1 | 50.7 | 25.5 |
| 56 | Janus-1.3B | CVPR2025 | Yes | 1.3B | 14.3 | 32.6 | 0 | 0 | 12.0 | 14.9 | 24.1 | 50.7 | 25.5 |
| 57 | UReader | EMNLP finding 2023 | Yes | 7B | 14.1 | 20.9 | 0 | 0 | 0 | 20.7 | 11.3 | 39.0 | 20.8 |
| 58 | LLaVAR | Arxiv 2023 | Yes | 13B | 12.4 | 13.8 | 0 | 0 | 8.3 | 15.2 | 4.4 | 42.4 | 15.0 |
| Rank | Method | Venue | Open-source | LLM Size | Average | Recognition | Extraction | Parsing | Understanding | Reasoning |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Gemini-2.5-Pro | No | 62.2 | 72.0 | 74.0 | 35.2 | 90.0 | 39.7 | ||
| 2 | Seed1.6-vision | No | 60.5 | 68.1 | 74.1 | 34.0 | 80.0 | 46.6 | ||
| 3 | Qwen3-Omni-30B-A3B-Instruct | Arxiv 2025 | Yes | 30B | 60.0 | 71.0 | 80.3 | 9 | 92.0 | 47.0 |
| 4 | MiniCPM-V-4.5-8B | Arxiv2025 | Yes | 8B | 58.8 | 70.1 | 70.1 | 32.5 | 92.0 | 29.4 |
| 5 | SAIL-VL2-8B | Arxiv2025 | Yes | 8B | 57.6 | 63.4 | 66.4 | 38.9 | 88.0 | 31.1 |
| 6 | Ovis2-8B | Yes | 7B | 56.0 | 61.0 | 67.7 | 43.6 | 82.0 | 25.6 | |
| 7 | WeThink-Qwen2.5VL-7B | Arxiv 2025 | Yes | 7B | 55.8 | 31.9 | 79.6 | 55.5 | 80.0 | 31.8 |
| 8 | Gemini1.5-Pro | Arxiv2024 | No | 55.5 | 71.4 | 63.8 | 30.5 | 82.0 | 29.9 | |
| 9 | Kimi-VL-A3B-16B | Arxiv 2025 | Yes | 16B | 54.1 | 54.0 | 71.1 | 32.5 | 84.0 | 28.7 |
| 10 | Step-1V | No | 53.4 | 65.2 | 64.9 | 33.1 | 78.0 | 25.5 | ||
| 11 | InterVL3-14B | Arxiv 2025 | Yes | 14B | 52.8 | 62.1 | 59.5 | 33.2 | 80.0 | 29.2 |
| 12 | InternVL3.5-14B | Arxiy2025 | Yes | 14B | 51.9 | 61.1 | 57.6 | 31.5 | 82.0 | 27.1 |
| 13 | GLM-4v-9B | Arxiy2024 | Yes | 9B | 51.7 | 60.6 | 65.2 | 32.4 | 82.0 | 18.2 |
| 14 | IntemVL3.5-8B | Arxiv 2025 | Yes | 8B | 50.3 | 62.4 | 58.7 | 28.8 | 72.0 | 29.6 |
| 15 | Qwen2.5-VL-7B | Arxiv2025 | Yes | 8B | 49.5 | 24.4 | 78.9 | 33.1 | 82.0 | 29.0 |
| 16 | InternVL3-8B | Yes | 8B | 49.0 | 57.7 | 55.8 | 29.9 | 72.0 | 29.4 | |
| 17 | Claude3.5-sonnet | No | 48.4 | 34.2 | 62.5 | 35.2 | 78.0 | 32.2 | ||
| 18 | DeepSeek-VL2-Small | Arxiv2024 | Yes | 16B | 48.1 | 51.6 | 56.3 | 27.8 | 79.6 | 25.3 |
| 19 | MiniCPM-V-2.6 | Arxiv2024 | Yes | 8B | 47.7 | 53.1 | 53.2 | 32.8 | 76.0 | 23.4 |
| 20 | MiniCPM-o-2.6 | Yes | 8B | 47.7 | 54.0 | 62.4 | 24.1 | 68.0 | 29.8 | |
| 21 | Claude-sonnet4-20250514 | No | 47.3 | 40.0 | 58.9 | 34.2 | 76.0 | 27.3 | ||
| 22 | GPT-5-2025-08-07 | No | 45.7 | 39.2 | 55.7 | 30.2 | 70.0 | 33.3 | ||
| 23 | GPT-4o | Arxiv2024 | No | 45.7 | 41.7 | 52.1 | 29.0 | 76.0 | 29.4 | |
| 24 | Qwen2-V-7B | Arxiy2024 | Yes | 8B | 44.7 | 23.7 | 63.5 | 27.9 | 80.0 | 28.5 |
| 25 | LLaVA-OneVision-1.5-8B-Instruct | Yes | 8B | 43.8 | 46.5 | 42.8 | 35.8 | 76.0 | 17.9 | |
| 26 | InternVL2.5-8B | Arxiv2024 | Yes | 8B | 42.8 | 42.8 | 47.9 | 27.3 | 80.0 | 23.5 |
| 27 | SAIL-VL-1.6-8B | Arxiv2025 | Yes | 8B | 42.6 | 35.8 | 41.5 | 35.7 | 76.0 | 23.9 |
| 28 | InternVL2.5-26B | Arxiy2024 | Yes | 20B | 41.9 | 40.2 | 42.7 | 25.6 | 74.0 | 27.0 |
| 29 | InterVL2-8B | SCIS 2024 | Yes | 8B | 41.3 | 35.2 | 42.8 | 26.1 | 78.0 | 24.4 |
| 30 | Liama-3.1-Nemotron-Nano-VL-8B-V1 | Yes | 8B | 40.1 | 38.2 | 54.9 | 26.6 | 66.0 | 14.8 | |
| 31 | IntermVL2-26B | SCIS 2024 | Yes | 20B | 38.1 | 20.4 | 50.7 | 29.0 | 76.0 | 14.5 |
| 32 | GPT-4o-mini | No | 37.4 | 20.0 | 53.6 | 27.9 | 66.0 | 19.6 | ||
| 33 | Phi4-MutiModal | Arxiv2025 | Yes | 5.6B | 37.3 | 30.5 | 40.5 | 42.7 | 56.0 | 16.9 |
| 34 | XComposer2-4KHD | NIPS 2025 | Yes | 8B | 32.4 | 12.9 | 38.6 | 37.5 | 60.0 | 13.1 |
| 35 | Ovis1.6-3B | Arxiv2024 | Yes | 3B | 31.7 | 22.5 | 33.3 | 31.5 | 54.0 | 17.0 |
| 36 | Grok4 | No | 22.7 | 8.1 | 33.3 | 16.1 | 40.0 | 16.1 | ||
| 37 | Monkey | CVPR2024 | Yes | 8B | 21.5 | 1.5 | 28.4 | 29.1 | 40.0 | 8.3 |
| 38 | TextMonkey | Arxiv2024 | Yes | 8B | 21.5 | 10.5 | 15.2 | 30.2 | 44.0 | 7.6 |
| 39 | Cambrian-1-8B | NIPS 2025 | Yes | 8B | 18.5 | 2.4 | 19.8 | 26.7 | 36.0 | 7.6 |
| 40 | LLaVA-OV-7B | Arxiy2024 | Yes | 8B | 17.4 | 5.4 | 13.6 | 20.3 | 34.0 | 13.6 |
| 41 | mPLUG-Ow3 | Arxiv2024 | Yes | 8B | 16.5 | 1.6 | 27.4 | 27.3 | 16.0 | 10.0 |
| 42 | Pixtral-12B | Arxiv2024 | Yes | 12B | 16.0 | 6.2 | 22.3 | 11.4 | 26.0 | 14.0 |
| 43 | Qwen-VL-chat | Arxiv2023 | Yes | 8B | 16.5 | 9.1 | 3.6 | 18.9 | 44.0 | 7.1 |
| 44 | Idefic3-8B | NeurlIPS 2024 Workshop | Yes | 8B | 15.6 | 2.9 | 29.0 | 12.3 | 26.0 | 7.9 |
| 45 | Qwen-VL | Arxiv 2023 | Yes | 8B | 15.6 | 4.3 | 0 | 30.6 | 38.0 | 5.1 |
| 46 | Molmo-7B | CVPR 2025 | Yes | 8B | 15.0 | 3.4 | 29.8 | 6.6 | 24.0 | 11.1 |
| 47 | DocOw2 | Arxiv2024 | Yes | 7B | 14.4 | 1.0 | 17.8 | 29.4 | 20.0 | 3.9 |
| 48 | Deepseek-VL-7B | Arxiv 2024 | Yes | 7B | 13.7 | 3.2 | 14.7 | 10.7 | 30.0 | 9.8 |
| 49 | CogVLM-chat | NIPS 2024 | Yes | 7B | 12.8 | 2.4 | 16.2 | 22.5 | 20.0 | 3.1 |
| 50 | Eagle-X5-7B | ICLR 2025 | Yes | 8B | 12.3 | 1.9 | 16.1 | 13.6 | 22.0 | 8.1 |
| 51 | VILA1.5-8B | CVPR2024 | Yes | 8B | 11.0 | 1.4 | 9.1 | 22.2 | 16.0 | 6.4 |
| 52 | Y-VL-6B | Arxiv 2024 | Yes | 6B | 10.4 | 1.6 | 6.4 | 28.8 | 10.0 | 5.3 |
| 53 | LLaVA-Nex-8B | Yes | 8B | 9.2 | 2.8 | 0.9 | 14.9 | 20.0 | 7.4 | |
| 54 | UReader | EMNLP finding 2023 | Yes | 7B | 9.0 | 0.3 | 2.0 | 28.1 | 12.0 | 2.4 |
| 55 | LLaVAR | Arxiv2023 | Yes | 13B | 8.6 | 2.2 | 2.0 | 27.1 | 10.0 | 1.9 |
| 56 | EMU2-chal | CVPR2024 | Yes | 37B | 8.2 | 1.2 | 3.0 | 29.3 | 4.0 | 3.6 |
| 57 | Janus-1.3B | CVPR 2025 | Yes | 1.3B | 7.5 | 4.1 | 2.2 | 10.4 | 14.0 | 6.7 |
以往针对LMMs的OCR评测基准主要关注基础的文字识别任务,但随着大模型的普及,现实世界的OCR需求已超越简单的“读出文字”。文档中的表格、图表、手写笔记、复杂的版式,文字图像的文本定位,以及基于文本的推理,都是LMMs面临的挑战。
现有评测基准大多任务单一、场景有限,导致模型得分迅速饱和,难以真实反映其在复杂应用中的能力。OCRBench v2提出一个综合全面的OCR测评基准,评估LMMs在面对文字定位、复杂场景理解与推理等广泛OCR任务时的真实表现。
OCRBench v2涵盖23种细分任务,涵盖了实际应用中常见的OCR需求场景。OCRBench v2将这些任务划分为8个核心能力维度:文本识别、文本定位、文本检测识别、关系抽取、元素解析、数学计算、视觉文本理解和知识推理,评价榜单中分别展示了模型在各个能力维度上的具体表现。
OCRBench v2的公开数据集包含来自80余个学术数据集及部分自有数据的1万条高质量QA,并经过人工审核,确保覆盖真实OCR应用中的多样化场景。
OCRBench v2还包括独立的私有数据,这部分数据包含人工采集并标注的1500条QA,其任务设置和场景覆盖范围均与公开数据保持一致。
实验发现,公开数据与私有数据榜单排名具有较高的一致性(详见论文分析),这证明了OCRBench v2任务设计、数据构造和评价指标的合理性,体现了其在衡量LMMs现有局限方面的重要价值。
闯红灯后停在路中间,会不会被扣分
RTX 5080游戏本实测DLSS
HarmonyOS免密认证方案
为何同样 25℃ ,湿度 95% 的空气里感觉闷热,而水里感觉冰冷?
让PPT表格变美观,三线表制作步骤详解
牙齿矫正疼不疼?
解决AutoCAD2026尺寸界线只显示一边的方法
赤坂丽(赤坂 麗/あかさか れい)电影全集
电脑无法开机,主板卡BOOT报错灯
1Password 密码管理器的最佳平替 Sticky Password
Excel2024绘制带指示线的同比柱形图
AutoCAD2026中多行文字双击后没有弹出「编辑工具栏」如何恢复?