--- dataset_info: - config_name: complexgrid_10x10 features: - name: image dtype: image - name: id dtype: string - name: caption dtype: string - name: answer dtype: string splits: - name: test num_bytes: 227929589200 num_examples: 50000 download_size: 215241037732 dataset_size: 227929589200 - config_name: complexgrid_3x3 features: - name: image dtype: image - name: id dtype: string - name: caption dtype: string - name: answer dtype: string splits: - name: test num_bytes: 1746798205.5 num_examples: 4500 download_size: 1742134215 dataset_size: 1746798205.5 - config_name: complexgrid_5x5 features: - name: image dtype: image - name: id dtype: string - name: caption dtype: string - name: answer dtype: string splits: - name: test num_bytes: 13902057100 num_examples: 12500 download_size: 13466500234 dataset_size: 13902057100 - config_name: complexgrid_7x7 features: - name: image dtype: image - name: id dtype: string - name: caption dtype: string - name: answer dtype: string splits: - name: test num_bytes: 54518960650 num_examples: 24500 download_size: 51624684839 dataset_size: 54518960650 - config_name: realworld_ArtBench features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 250211674 num_examples: 44 - name: validation num_bytes: 131797353 num_examples: 30 - name: test num_bytes: 1708776579 num_examples: 232 download_size: 2090852741 dataset_size: 2090785606 - config_name: realworld_Autonomous_Driving features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 19903320 num_examples: 36 - name: validation num_bytes: 17766200 num_examples: 30 - name: test num_bytes: 135505593 num_examples: 234 download_size: 159545381 dataset_size: 173175113 - config_name: realworld_CAD features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 16459800 num_examples: 46 - name: validation num_bytes: 9915669 num_examples: 30 - name: test num_bytes: 70196437 num_examples: 221 download_size: 94265399 dataset_size: 96571906 - config_name: realworld_Diagram_and_Table features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 13028678 num_examples: 39 - name: validation num_bytes: 10033883 num_examples: 30 - name: test num_bytes: 78307813 num_examples: 231 download_size: 79562780 dataset_size: 101370374 - config_name: realworld_DocStruct4M features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 20720911 num_examples: 42 - name: validation num_bytes: 11993873 num_examples: 30 - name: test num_bytes: 101409675 num_examples: 224 download_size: 119403891 dataset_size: 134124459 - config_name: realworld_Galaxy features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 1388060569 num_examples: 30 - name: validation num_bytes: 2141262985 num_examples: 30 - name: test num_bytes: 1870319322 num_examples: 27 download_size: 5389481215 dataset_size: 5399642876 - config_name: realworld_Grass features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 90972563 num_examples: 36 - name: validation num_bytes: 76612991 num_examples: 30 - name: test num_bytes: 587887567 num_examples: 234 download_size: 755431650 dataset_size: 755473121 - config_name: realworld_HR-Bench features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 156586272 num_examples: 51 - name: validation num_bytes: 90110107 num_examples: 30 - name: test num_bytes: 983783374 num_examples: 316 download_size: 1090513862 dataset_size: 1230479753 - config_name: realworld_HRIQ features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 96290107 num_examples: 47 - name: validation num_bytes: 56753956 num_examples: 30 - name: test num_bytes: 461250556 num_examples: 223 download_size: 614093072 dataset_size: 614294619 - config_name: realworld_HRVQA features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 65286480 num_examples: 36 - name: validation num_bytes: 54826810 num_examples: 30 - name: test num_bytes: 368450636 num_examples: 207 download_size: 484937958 dataset_size: 488563926 - config_name: realworld_InfographicVQA features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 28268627 num_examples: 42 - name: validation num_bytes: 20823588 num_examples: 30 - name: test num_bytes: 165275830 num_examples: 228 download_size: 142296366 dataset_size: 214368045 - config_name: realworld_Izembek features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 507979769 num_examples: 37 - name: validation num_bytes: 411223282 num_examples: 30 - name: test num_bytes: 3294031877 num_examples: 233 download_size: 4211319748 dataset_size: 4213234928 - config_name: realworld_LungHist700 features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 37181374 num_examples: 37 - name: validation num_bytes: 29542107 num_examples: 30 - name: test num_bytes: 242748422 num_examples: 241 download_size: 309370044 dataset_size: 309471903 - config_name: realworld_MAME features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 357833889 num_examples: 46 - name: validation num_bytes: 290101548 num_examples: 30 - name: test num_bytes: 1279306240 num_examples: 224 download_size: 1926012458 dataset_size: 1927241677 - config_name: realworld_MMAD features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 47925120 num_examples: 39 - name: validation num_bytes: 42782274 num_examples: 30 - name: test num_bytes: 353884625 num_examples: 231 download_size: 425912773 dataset_size: 444592019 - config_name: realworld_MileBench features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 34531401 num_examples: 43 - name: validation num_bytes: 18356281 num_examples: 30 - name: test num_bytes: 165377058 num_examples: 227 download_size: 214845061 dataset_size: 218264740 - config_name: realworld_MuirBench features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 25773649 num_examples: 34 - name: validation num_bytes: 17765807 num_examples: 30 - name: test num_bytes: 134118072 num_examples: 236 download_size: 149615394 dataset_size: 177657528 - config_name: realworld_NovaChart features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 4573600 num_examples: 38 - name: validation num_bytes: 3754212 num_examples: 30 - name: test num_bytes: 28228274 num_examples: 229 download_size: 31492431 dataset_size: 36556086 - config_name: realworld_OCR_in_the_Wild features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 28707630 num_examples: 45 - name: validation num_bytes: 24214125 num_examples: 30 - name: test num_bytes: 173337691 num_examples: 225 download_size: 222693503 dataset_size: 226259446 - config_name: realworld_PANDA features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 807349639 num_examples: 44 - name: validation num_bytes: 531740280 num_examples: 30 - name: test num_bytes: 4167522801 num_examples: 226 download_size: 5187495009 dataset_size: 5506612720 - config_name: realworld_Remote_Sensing features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 223816278 num_examples: 43 - name: validation num_bytes: 141777098 num_examples: 30 - name: test num_bytes: 1137375998 num_examples: 227 download_size: 1495392037 dataset_size: 1502969374 - config_name: realworld_VQA-RAD features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 3415345 num_examples: 37 - name: validation num_bytes: 2640198 num_examples: 30 - name: test num_bytes: 14184726 num_examples: 158 download_size: 9760010 dataset_size: 20240269 - config_name: realworld_VStar_Bench features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 26804623 num_examples: 38 - name: validation num_bytes: 19355778 num_examples: 30 - name: test num_bytes: 118319180 num_examples: 164 download_size: 164398648 dataset_size: 164479581 - config_name: realworld_Video_Monitoring features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 17047821 num_examples: 38 - name: validation num_bytes: 13851903 num_examples: 30 - name: test num_bytes: 109345706 num_examples: 232 download_size: 131835084 dataset_size: 140245430 - config_name: realworld_VisDiffBench features: - name: id dtype: int64 - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 51943569 num_examples: 32 - name: validation num_bytes: 47266535 num_examples: 30 - name: test num_bytes: 135552775 num_examples: 88 download_size: 233255572 dataset_size: 234762879 - config_name: realworld_combined features: - name: id dtype: string - name: image dtype: image - name: question dtype: string - name: answer dtype: string splits: - name: testmini num_bytes: 4320671593 num_examples: 1000 - name: validation num_bytes: 4216267981 num_examples: 750 - name: test num_bytes: 8666836592.13 num_examples: 5318 download_size: 25695272223 dataset_size: 17203776166.129997 - config_name: whitebackground_10x10 features: - name: image dtype: image - name: id dtype: int64 - name: question dtype: string - name: answer sequence: string splits: - name: test num_bytes: 17587646750 num_examples: 50000 download_size: 2631010727 dataset_size: 17587646750 - config_name: whitebackground_1x1 features: - name: image dtype: image - name: id dtype: int64 - name: question dtype: string - name: answer sequence: string splits: - name: test num_bytes: 18170859 num_examples: 500 download_size: 18090243 dataset_size: 18170859 - config_name: whitebackground_3x3 features: - name: image dtype: image - name: id dtype: int64 - name: question dtype: string - name: answer sequence: string splits: - name: test num_bytes: 278404807.5 num_examples: 4500 download_size: 167651986 dataset_size: 278404807.5 - config_name: whitebackground_5x5 features: - name: image dtype: image - name: id dtype: int64 - name: question dtype: string - name: answer sequence: string splits: - name: test num_bytes: 1424175987.5 num_examples: 12500 download_size: 518486585 dataset_size: 1424175987.5 - config_name: whitebackground_7x7 features: - name: image dtype: image - name: id dtype: int64 - name: question dtype: string - name: answer sequence: string splits: - name: test num_bytes: 4643563645 num_examples: 24500 download_size: 1103821117 dataset_size: 4643563645 configs: - config_name: complexgrid_10x10 data_files: - split: test path: complexgrid_10x10/test-* - config_name: complexgrid_3x3 data_files: - split: test path: complexgrid_3x3/test-* - config_name: complexgrid_5x5 data_files: - split: test path: complexgrid_5x5/test-* - config_name: complexgrid_7x7 data_files: - split: test path: complexgrid_7x7/test-* - config_name: realworld_ArtBench data_files: - split: testmini path: realworld_ArtBench/train-* - split: validation path: realworld_ArtBench/validation-* - split: test path: realworld_ArtBench/test-* - config_name: realworld_Autonomous_Driving data_files: - split: testmini path: realworld_Autonomous_Driving/train-* - split: validation path: realworld_Autonomous_Driving/validation-* - split: test path: realworld_Autonomous_Driving/test-* - config_name: realworld_CAD data_files: - split: testmini path: realworld_CAD/train-* - split: validation path: realworld_CAD/validation-* - split: test path: realworld_CAD/test-* - config_name: realworld_Diagram_and_Table data_files: - split: testmini path: realworld_Diagram_and_Table/train-* - split: validation path: realworld_Diagram_and_Table/validation-* - split: test path: realworld_Diagram_and_Table/test-* - config_name: realworld_DocStruct4M data_files: - split: testmini path: realworld_DocStruct4M/train-* - split: validation path: realworld_DocStruct4M/validation-* - split: test path: realworld_DocStruct4M/test-* - config_name: realworld_Galaxy data_files: - split: testmini path: realworld_Galaxy/train-* - split: validation path: realworld_Galaxy/validation-* - split: test path: realworld_Galaxy/test-* - config_name: realworld_Grass data_files: - split: testmini path: realworld_Grass/train-* - split: validation path: realworld_Grass/validation-* - split: test path: realworld_Grass/test-* - config_name: realworld_HR-Bench data_files: - split: testmini path: realworld_HR-Bench/train-* - split: validation path: realworld_HR-Bench/validation-* - split: test path: realworld_HR-Bench/test-* - config_name: realworld_HRIQ data_files: - split: testmini path: realworld_HRIQ/train-* - split: validation path: realworld_HRIQ/validation-* - split: test path: realworld_HRIQ/test-* - config_name: realworld_HRVQA data_files: - split: testmini path: realworld_HRVQA/train-* - split: validation path: realworld_HRVQA/validation-* - split: test path: realworld_HRVQA/test-* - config_name: realworld_InfographicVQA data_files: - split: testmini path: realworld_InfographicVQA/train-* - split: validation path: realworld_InfographicVQA/validation-* - split: test path: realworld_InfographicVQA/test-* - config_name: realworld_Izembek data_files: - split: testmini path: realworld_Izembek/train-* - split: validation path: realworld_Izembek/validation-* - split: test path: realworld_Izembek/test-* - config_name: realworld_LungHist700 data_files: - split: testmini path: realworld_LungHist700/train-* - split: validation path: realworld_LungHist700/validation-* - split: test path: realworld_LungHist700/test-* - config_name: realworld_MAME data_files: - split: testmini path: realworld_MAME/train-* - split: validation path: realworld_MAME/validation-* - split: test path: realworld_MAME/test-* - config_name: realworld_MMAD data_files: - split: testmini path: realworld_MMAD/train-* - split: validation path: realworld_MMAD/validation-* - split: test path: realworld_MMAD/test-* - config_name: realworld_MileBench data_files: - split: testmini path: realworld_MileBench/train-* - split: validation path: realworld_MileBench/validation-* - split: test path: realworld_MileBench/test-* - config_name: realworld_MuirBench data_files: - split: testmini path: realworld_MuirBench/train-* - split: validation path: realworld_MuirBench/validation-* - split: test path: realworld_MuirBench/test-* - config_name: realworld_NovaChart data_files: - split: testmini path: realworld_NovaChart/train-* - split: validation path: realworld_NovaChart/validation-* - split: test path: realworld_NovaChart/test-* - config_name: realworld_OCR_in_the_Wild data_files: - split: testmini path: realworld_OCR_in_the_Wild/train-* - split: validation path: realworld_OCR_in_the_Wild/validation-* - split: test path: realworld_OCR_in_the_Wild/test-* - config_name: realworld_PANDA data_files: - split: testmini path: realworld_PANDA/train-* - split: validation path: realworld_PANDA/validation-* - split: test path: realworld_PANDA/test-* - config_name: realworld_Remote_Sensing data_files: - split: testmini path: realworld_Remote_Sensing/train-* - split: validation path: realworld_Remote_Sensing/validation-* - split: test path: realworld_Remote_Sensing/test-* - config_name: realworld_VQA-RAD data_files: - split: testmini path: realworld_VQA-RAD/train-* - split: validation path: realworld_VQA-RAD/validation-* - split: test path: realworld_VQA-RAD/test-* - config_name: realworld_VStar_Bench data_files: - split: testmini path: realworld_VStar_Bench/train-* - split: validation path: realworld_VStar_Bench/validation-* - split: test path: realworld_VStar_Bench/test-* - config_name: realworld_Video_Monitoring data_files: - split: testmini path: realworld_Video_Monitoring/train-* - split: validation path: realworld_Video_Monitoring/validation-* - split: test path: realworld_Video_Monitoring/test-* - config_name: realworld_VisDiffBench data_files: - split: testmini path: realworld_VisDiffBench/train-* - split: validation path: realworld_VisDiffBench/validation-* - split: test path: realworld_VisDiffBench/test-* - config_name: realworld_combined data_files: - split: testmini path: realworld_combined/train-* - split: validation path: realworld_combined/validation-* - split: test path: realworld_combined/test-* - config_name: whitebackground_10x10 data_files: - split: test path: whitebackground_10x10/test-* - config_name: whitebackground_1x1 data_files: - split: test path: whitebackground_1x1/test-* - config_name: whitebackground_3x3 data_files: - split: test path: whitebackground_3x3/test-* - config_name: whitebackground_5x5 data_files: - split: test path: whitebackground_5x5/test-* - config_name: whitebackground_7x7 data_files: - split: test path: whitebackground_7x7/test-* --- #
HRScene - High Resolution Image Understanding
🌐 Homepage | 🤗 Dataset | 📖 arXiv | GitHub ## ⭐ About HRScene

We introduce HRScene, a novel unified benchmark for HRI understanding with rich scenes. HRScene incorporates 25 real-world datasets and 2 synthetic diagnostic datasets with resolutions ranging from 1,024 × 1,024 to 35,503 × 26,627. HRScene is collected and re-annotated by 10 graduate-level annotators, covering 25 scenarios, ranging from microscopic and radiology images to street views, long-range pictures, and telescope images. It includes high-resolution images of real-world objects, scanned documents, and composite multi-image.

Dataset Sample Image

Some examples of HRScene. Blue ones are diagnostic datasets, and purple ones are real-world datasets.

HRScene consists of 7,073 samples, divided into three splits:

Val contains 750 samples. These samples are identical to human-annotated ones, designed for fine-grained validation of the users' VLM settings. Testmini comprises 1,000 samples, picked from each HRScene real-world dataset, intended for rapid model development evaluation or for those with limited computing resources. Test features the remaining 5,323 samples for standard evaluation. Notably, the answer labels for test will not be publicly released to facilitate fair evaluation. Instead, we maintain an online evaluation platform for user submissions.


## 📖 Dataset Usage ### Data Downloading By using our [pipeline](https://github.com/psunlpgroup/HRScene), you dont need to download the dataset manually. For whitebackground and complexgrid, you only need to set the `dataset_name` for the tester like we did in the [Diagnosis/example.py](https://github.com/psunlpgroup/HRScene/blob/main/Diagnosis/example.py). ```python tester = DiagnosisTester(model=model, dataset_name="complexgrid_3x3", num_samples=150) ``` For realworld, you need to set the `dataset_name` and `split` for the tester like we did in the [RealWorld/example.py](https://github.com/psunlpgroup/HRScene/blob/main/RealWorld/example.py). ```python tester = RealWorldTester(model=model, dataset_name="realworld_combined", split="test") ``` Or you wanna download the dataset manually, you can use the following code: ```python from datasets import load_dataset # for whitebackground and complexgrid, we only have 'test' split dataset = load_dataset("Wenliang04/HRScene", "whitebackground_1x1") for sample in dataset['test']: print(sample) # for realworld, we have 'testmini', 'validation', 'test' splits dataset = load_dataset("Wenliang04/HRScene", "realworld_combined") for sample in dataset['test']: print(sample) ``` ### Data Format #### WhiteBackground id: int, image: PIL.JpegImagePlugin.JpegImageFile, question: str, answer: list[str] ```bash {'image': , 'id': 0, 'question': 'Is it daytime?', 'answer': ['no', 'no', 'no', 'no', 'no', 'no', 'no', 'no', 'no', 'no']} ``` #### ComplexGrid id: str, image: PIL.JpegImagePlugin.JpegImageFile, caption: str, answer: str ```bash {'image': , 'id': '0_0_0', 'caption': 'A nice living room has chairs and a love seat.', 'answer': 'row: 1, col: 1'} ``` #### RealWorld id: int, image: PIL.Image.Image, question: str, answer: str ```bash {'id': 0, 'image': , 'question': 'What is motion of the pedestrian wearing blue top on the left?\n(A) crossing the crosswalk\n(B) standing\n(C) jaywalking (illegally crossing not at pedestrian crossing)\n(D) walking on the sidewalk\n(E) The image does not feature the object', 'answer': 'None'} ``` ## 🏆 Leaderboard 🏆 ### Leaderboard on the RealWorld Task on the Test Split (Show Top 5 Only) | **#** | **Model** | **Art** | **Daily** | **Medical** | **Paper** | **Remote** | **Research** | **Sub-Img** | **Urban** | **Average** | | ----- | ------------------------ | ------- | --------- | ----------- | --------- | ---------- | ------------ | ----------- | --------- | ----------- | | 1 | **Qwen2-72B-Instruct** | 75.9 | 66.2 | 43.7 | 78.1 | 52.5 | 39.4 | 74.9 | 44.7 | **61.9** | | 2 | **gemini-2-flash** | 76.5 | 62.3 | 51.9 | 75.1 | 47.6 | 34.9 | 68.6 | 44.5 | **59.8** | | 3 | **InternVL2-40B** | 74.4 | 62.7 | 38.1 | 70.9 | 44.2 | 43.2 | 74.1 | 44.4 | **58.5** | | 4 | **Llava-OneVision-72B** | 68.3 | 64.6 | 42.0 | 68.9 | 46.2 | 52.1 | 68.7 | 40.3 | **57.5** | | 5 | **Qwen2-VL-7B-Instruct** | 69.5 | 64.2 | 40.4 | 64.6 | 50.6 | 36.7 | 71.4 | 40.2 | **56.7** | We provide a simple pipeline for automatic model prediction and submission file geneartion! You can find pipeline from our [Github](https://github.com/psunlpgroup/HRScene/tree/main) under "🔮 Evaluations on HRScene for RealWorld Task" section ## ✅ Cite ```bibtex @article{zhang2025hrscene, title={HRScene: How Far Are VLMs from Effective High-Resolution Image Understanding?}, author={Zhang, Yusen and Zheng, Wenliang and Madasu, Aashrith and Shi, Peng and Kamoi, Ryo and Zhou, Hao and Zou, Zhuoyang and Zhao, Shu and Das, Sarkar Snigdha Sarathi and Gupta, Vipul and Lu, Xiaoxin and Zhang, Nan and Zhang, Ranran Haoran and Iyer, Avitej and Lou, Renze and Yin, Wenpeng and Zhang, Rui}, journal={arXiv preprint}, year={2025} } ```