Datasets:
image imagewidth (px) 160 5.39k | sample_id stringlengths 31 50 | question stringlengths 19 193 | answer stringclasses 108
values | split stringclasses 20
values | image_height int64 120 5.76k ⌀ | image_width int64 160 5.39k ⌀ |
|---|---|---|---|---|---|---|
holocount_atomic_counting_00001 | How many giraffe are there in the image? | 1 | Atomic Counting | 343 | 500 | |
holocount_atomic_counting_00002 | How many cat are there in the image? | 1 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00003 | How many ski_boot are there in the image? | 2 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00004 | How many polo_shirt are there in the image? | 1 | Atomic Counting | 640 | 427 | |
holocount_atomic_counting_00005 | How many trousers are there in the image? | 3 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00006 | How many wine_bottle are there in the image? | 3 | Atomic Counting | 640 | 386 | |
holocount_atomic_counting_00007 | How many elephant are there in the image? | 5 | Atomic Counting | 428 | 640 | |
holocount_atomic_counting_00008 | How many pineapple are there in the image? | 5 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00009 | How many shirt are there in the image? | 1 | Atomic Counting | 338 | 500 | |
holocount_atomic_counting_00010 | How many helmet are there in the image? | 6 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00011 | How many cowboy_hat are there in the image? | 1 | Atomic Counting | 427 | 640 | |
holocount_atomic_counting_00012 | How many umbrella are there in the image? | 5 | Atomic Counting | 427 | 640 | |
holocount_atomic_counting_00013 | How many mug are there in the image? | 4 | Atomic Counting | 612 | 612 | |
holocount_atomic_counting_00014 | How many egg are there in the image? | 12 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00015 | How many muffin are there in the image? | 1 | Atomic Counting | 160 | 288 | |
holocount_atomic_counting_00016 | How many scissors are there in the image? | 12 | Atomic Counting | 481 | 640 | |
holocount_atomic_counting_00017 | How many coaster are there in the image? | 1 | Atomic Counting | 640 | 480 | |
holocount_atomic_counting_00018 | How many laptop_computer are there in the image? | 1 | Atomic Counting | 424 | 640 | |
holocount_atomic_counting_00019 | How many poster are there in the image? | 1 | Atomic Counting | 640 | 480 | |
holocount_atomic_counting_00020 | How many clock_tower are there in the image? | 1 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00021 | How many toothbrush are there in the image? | 1 | Atomic Counting | 640 | 427 | |
holocount_atomic_counting_00022 | How many elephant are there in the image? | 5 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00023 | How many plate are there in the image? | 12 | Atomic Counting | 640 | 431 | |
holocount_atomic_counting_00024 | How many beer_bottle are there in the image? | 7 | Atomic Counting | 640 | 480 | |
holocount_atomic_counting_00025 | How many snowboard are there in the image? | 1 | Atomic Counting | 500 | 332 | |
holocount_atomic_counting_00026 | How many knife are there in the image? | 1 | Atomic Counting | 612 | 612 | |
holocount_atomic_counting_00027 | How many helmet are there in the image? | 1 | Atomic Counting | 426 | 640 | |
holocount_atomic_counting_00028 | How many suitcase are there in the image? | 12 | Atomic Counting | 372 | 640 | |
holocount_atomic_counting_00029 | How many sandwich are there in the image? | 1 | Atomic Counting | 481 | 640 | |
holocount_atomic_counting_00030 | How many baseball_bat are there in the image? | 1 | Atomic Counting | 640 | 426 | |
holocount_atomic_counting_00031 | How many boat are there in the image? | 4 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00032 | How many baseball_bat are there in the image? | 1 | Atomic Counting | 427 | 640 | |
holocount_atomic_counting_00033 | How many person are there in the image? | 2 | Atomic Counting | 640 | 512 | |
holocount_atomic_counting_00034 | How many teddy_bear are there in the image? | 1 | Atomic Counting | 332 | 500 | |
holocount_atomic_counting_00035 | How many umbrella are there in the image? | 11 | Atomic Counting | 375 | 500 | |
holocount_atomic_counting_00036 | How many chair are there in the image? | 1 | Atomic Counting | 640 | 427 | |
holocount_atomic_counting_00037 | How many teddy_bear are there in the image? | 15 | Atomic Counting | 477 | 640 | |
holocount_atomic_counting_00038 | How many bathtub are there in the image? | 1 | Atomic Counting | 640 | 424 | |
holocount_atomic_counting_00039 | How many lantern are there in the image? | 1 | Atomic Counting | 640 | 640 | |
holocount_atomic_counting_00040 | How many fireplug are there in the image? | 1 | Atomic Counting | 640 | 480 | |
holocount_atomic_counting_00041 | How many skirt are there in the image? | 1 | Atomic Counting | 427 | 640 | |
holocount_atomic_counting_00042 | How many vase are there in the image? | 1 | Atomic Counting | 500 | 441 | |
holocount_atomic_counting_00043 | How many bus_(vehicle) are there in the image? | 2 | Atomic Counting | 425 | 640 | |
holocount_atomic_counting_00044 | How many bathtub are there in the image? | 1 | Atomic Counting | 640 | 427 | |
holocount_atomic_counting_00045 | How many tennis_racket are there in the image? | 2 | Atomic Counting | 640 | 427 | |
holocount_atomic_counting_00046 | How many trophy_cup are there in the image? | 8 | Atomic Counting | 615 | 640 | |
holocount_atomic_counting_00047 | How many mirror are there in the image? | 1 | Atomic Counting | 483 | 640 | |
holocount_atomic_counting_00048 | How many suitcase are there in the image? | 3 | Atomic Counting | 375 | 500 | |
holocount_atomic_counting_00049 | How many fireplug are there in the image? | 1 | Atomic Counting | 640 | 480 | |
holocount_atomic_counting_00050 | How many person are there in the image? | 1 | Atomic Counting | 426 | 640 | |
holocount_atomic_counting_00051 | How many banana are there in the image? | 1 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00052 | How many dress are there in the image? | 1 | Atomic Counting | 640 | 640 | |
holocount_atomic_counting_00053 | How many bird are there in the image? | 13 | Atomic Counting | 640 | 446 | |
holocount_atomic_counting_00054 | How many necktie are there in the image? | 1 | Atomic Counting | 478 | 640 | |
holocount_atomic_counting_00055 | How many skateboard are there in the image? | 12 | Atomic Counting | 612 | 612 | |
holocount_atomic_counting_00056 | How many surfboard are there in the image? | 1 | Atomic Counting | 399 | 640 | |
holocount_atomic_counting_00057 | How many dinghy are there in the image? | 1 | Atomic Counting | 640 | 480 | |
holocount_atomic_counting_00058 | How many glass_(drink_container) are there in the image? | 3 | Atomic Counting | 640 | 478 | |
holocount_atomic_counting_00059 | How many airplane are there in the image? | 3 | Atomic Counting | 425 | 640 | |
holocount_atomic_counting_00060 | How many tomato are there in the image? | 6 | Atomic Counting | 424 | 640 | |
holocount_atomic_counting_00061 | How many cupcake are there in the image? | 14 | Atomic Counting | 483 | 640 | |
holocount_atomic_counting_00062 | How many pigeon are there in the image? | 2 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00063 | How many statue_(sculpture) are there in the image? | 10 | Atomic Counting | 458 | 640 | |
holocount_atomic_counting_00064 | How many doughnut are there in the image? | 12 | Atomic Counting | 612 | 612 | |
holocount_atomic_counting_00065 | How many computer_keyboard are there in the image? | 3 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00066 | How many bench are there in the image? | 1 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00067 | How many cupcake are there in the image? | 4 | Atomic Counting | 612 | 612 | |
holocount_atomic_counting_00068 | How many cistern are there in the image? | 1 | Atomic Counting | 428 | 640 | |
holocount_atomic_counting_00069 | How many bird are there in the image? | 15 | Atomic Counting | 640 | 449 | |
holocount_atomic_counting_00070 | How many frisbee are there in the image? | 1 | Atomic Counting | 415 | 640 | |
holocount_atomic_counting_00071 | How many bird are there in the image? | 1 | Atomic Counting | 424 | 640 | |
holocount_atomic_counting_00072 | How many apple slices are there in the image? | 5 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00073 | How many sock are there in the image? | 2 | Atomic Counting | 457 | 640 | |
holocount_atomic_counting_00074 | How many toilet are there in the image? | 1 | Atomic Counting | 500 | 375 | |
holocount_atomic_counting_00075 | How many flag are there in the image? | 7 | Atomic Counting | 427 | 640 | |
holocount_atomic_counting_00076 | How many laptop_computer are there in the image? | 1 | Atomic Counting | 428 | 640 | |
holocount_atomic_counting_00077 | How many lampshade are there in the image? | 2 | Atomic Counting | 444 | 640 | |
holocount_atomic_counting_00078 | How many fork are there in the image? | 1 | Atomic Counting | 612 | 612 | |
holocount_atomic_counting_00079 | How many doughnut are there in the image? | 20 | Atomic Counting | 425 | 640 | |
holocount_atomic_counting_00080 | How many pop_(soda) are there in the image? | 1 | Atomic Counting | 428 | 640 | |
holocount_atomic_counting_00081 | How many carrot are there in the image? | 10 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00082 | How many person are there in the image? | 6 | Atomic Counting | 612 | 612 | |
holocount_atomic_counting_00083 | How many poster are there in the image? | 9 | Atomic Counting | 427 | 640 | |
holocount_atomic_counting_00084 | How many glass_(drink_container) are there in the image? | 4 | Atomic Counting | 428 | 640 | |
holocount_atomic_counting_00085 | How many saddle_blanket are there in the image? | 1 | Atomic Counting | 400 | 600 | |
holocount_atomic_counting_00086 | How many painting are there in the image? | 4 | Atomic Counting | 500 | 375 | |
holocount_atomic_counting_00087 | How many sink are there in the image? | 1 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00088 | How many wet_suit are there in the image? | 1 | Atomic Counting | 427 | 640 | |
holocount_atomic_counting_00089 | How many minivan are there in the image? | 5 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00090 | How many television_set are there in the image? | 1 | Atomic Counting | 640 | 478 | |
holocount_atomic_counting_00091 | How many teddy_bear are there in the image? | 3 | Atomic Counting | 612 | 612 | |
holocount_atomic_counting_00092 | How many banana are there in the image? | 7 | Atomic Counting | 543 | 640 | |
holocount_atomic_counting_00093 | How many grizzly are there in the image? | 3 | Atomic Counting | 425 | 640 | |
holocount_atomic_counting_00094 | How many plate are there in the image? | 6 | Atomic Counting | 480 | 640 | |
holocount_atomic_counting_00095 | How many boat are there in the image? | 4 | Atomic Counting | 424 | 640 | |
holocount_atomic_counting_00096 | How many pizza are there in the image? | 2 | Atomic Counting | 478 | 640 | |
holocount_atomic_counting_00097 | How many umbrella are there in the image? | 1 | Atomic Counting | 427 | 640 | |
holocount_atomic_counting_00098 | How many bus_(vehicle) are there in the image? | 1 | Atomic Counting | 375 | 500 | |
holocount_atomic_counting_00099 | How many printer are there in the image? | 1 | Atomic Counting | 333 | 500 | |
holocount_atomic_counting_00100 | How many refrigerator are there in the image? | 1 | Atomic Counting | 427 | 640 |
HoloCount: A Holistic Visual Counting Benchmark for MLLMs
Abstract
Visual counting is a fundamental pillar of multimodal intelligence, requiring a seamless integration of fine-grained grounding and spatial reasoning. While Multimodal Large Language Models (MLLMs) have achieved remarkable success in qualitative scene understanding, their quantitative precision remains a significant bottleneck, often characterized by persistent numerical hallucinations. Existing counting benchmarks primarily focus on basic perception in simplified contexts, failing to capture the complex failure modes that emerge under logical constraints or adversarial conditions. To address these limitations, we introduce HoloCount, a holistic and diagnostically rich benchmark structured around a three-level hierarchical taxonomy. HoloCount evaluates MLLMs across: (1) Semantic Counting, focusing on atomic and property-based enumeration; (2) Analytical Counting, assessing logical composition through spatial and set-based reasoning; and (3) Robustness Testing, probing model integrity against adverse scenarios and grounded counter-priors, such as high-density scenes and linguistic biases. Through an exhaustive evaluation of over 20 state-of-the-art MLLMs, we reveal a critical performance gap: even top-tier models degrade significantly as tasks transition from perception to complex analytical reasoning and adverse scenarios. Our findings provide a systematic landscape of current MLLM counting capabilities and offer a roadmap for developing more grounded and reliable multimodal systems.
Overview
- 2,480 QA pairs
- 20 fine-grained subsets
- 1,481 visual concepts
- 30+ evaluated MLLMs
Taxonomy
The benchmark features a three-level hierarchical taxonomy:
- Semantic Counting — Atomic and property-based enumeration (6 subsets)
- Analytical Counting — Logical composition through spatial and set-based reasoning (7 subsets)
- Robustness Testing — Model integrity against adverse scenarios (7 subsets)
Dataset Composition
| Split | Count |
|---|---|
| Action & State Attribution | 102 |
| Aerial Perspective Shift | 109 |
| Atomic Counting | 182 |
| Categorical Cardinality | 101 |
| Chromatic Attribution | 202 |
| Complementary Exclusion | 101 |
| Coordinate-Prompt Region Grounding | 84 |
| Differential Comparison | 100 |
| General Semantic Attribution | 101 |
| High-Density Enumeration | 100 |
| Joint-Set Aggregation | 100 |
| Linguistic Prior Conflict | 163 |
| Material Attribution | 101 |
| Null-Target Prompting | 250 |
| Partial Object Occlusion | 103 |
| Relative Canonical Orientation | 78 |
| Scale Attribution | 100 |
| Small-Scale Enumeration | 101 |
| Visual Distractor Illusion | 200 |
| Visual-Prompt Region Grounding | 102 |
Usage
from datasets import load_dataset
dataset = load_dataset("MM-MVR/HoloCount")
Fields
image: the image filequestion: counting questionanswer: ground truth count (string)split: capability categorysample_id: unique identifierimage_height: original image height in pixelsimage_width: original image width in pixels
Citation
@article{deng2026holocount,
title={HoloCount: A Holistic Visual Counting Benchmark for MLLMs},
author={Deng, Jinhong and Qiao, Limeng and Wan, Guanglu},
year={2026}
}
License
This dataset is released under the CC BY 4.0 License.
Contact
- Jinhong Deng (dengjinhong@meituan.com)
- Downloads last month
- 344