---
pretty_name: CDBench
license: cc-by-nc-sa-4.0
language:
- en
task_categories:
- visual-question-answering
- question-answering
size_categories:
- 10K<n<100K
tags:
- change-detection
- multimodal
- vision-language
- benchmark
- remote-sensing
- industrial-inspection
- image
- text
configs:
- config_name: viewer
  data_files:
  - split: train
    path: questions.csv
---

# CDBench: A Comprehensive Multimodal Dataset and Evaluation Benchmark for General Change Detection

**16,086 image pairs · 73,122 multiple-choice QA items · 7 tasks · 3 domains**

[Project website](http://mm-cd.org) | [Dataset files](https://huggingface.co/datasets/MM-CD/CDBench/tree/main) | [Questions and feedback](https://huggingface.co/datasets/MM-CD/CDBench/discussions)

CDBench evaluates how multimodal large language models (MLLMs) understand images and reason about changes across **remote sensing, industrial inspection, and commodity products**. It connects image content recognition with change discrimination, regional localization, semantic classification, description, and plausible cause inference through a shared multiple-choice question (MCQ) format.

The accompanying **ChangeAgent** framework combines an MLLM with expert visual guidance and retrieval-augmented generation. In the reported evaluation, it achieves **78.25%** average accuracy across Q1–Q7, compared with **71.19%** for Gemini 1.5 Pro and **70.68%** for GPT-4o.

## Dataset at a glance

The counts below are computed directly from the released `questions.csv`. An image pair is a unique `(target_image, reference_image)` tuple; a QA item is one CSV row.

| Domain | Source dataset | Image pairs | QA items | Task coverage |
|---|---|---:|---:|---|
| Remote sensing | CDD | 2,343 | 9,468 | Q1–Q4, Q6–Q7 |
| Remote sensing | LEVIR-CD | 3,093 | 12,444 | Q1–Q4, Q6–Q7 |
| Remote sensing | SYSU-CD | 2,316 | 9,409 | Q1–Q4, Q6–Q7 |
| Industrial inspection | MVTec-AD | 1,685 | 9,442 | Q1–Q7 |
| Industrial inspection | MVTec-LOCO | 1,557 | 8,016 | Q1–Q7 |
| Industrial inspection | VisA | 2,132 | 10,189 | Q1–Q7 |
| Commodity products | GoodsAD | 2,960 | 14,154 | Q1–Q7 |
| **Total** | **7 source datasets** | **16,086** | **73,122** | **7 tasks overall** |

Not every image pair has all seven questions. The released remote-sensing subsets contain no Q5 items, and coverage varies across tasks and samples. The totals describe the released annotations, not the sizes of the original source datasets.

## Seven evaluation tasks

| ID | Task | What the model selects | QA items |
|---|---|---|---:|
| Q1 | Image Content Classification | The primary scene type or dominant content category of one image. | 16,075 |
| Q2 | Image Content Description | The description that best matches objects, attributes, composition, and spatial relations in one image. | 8,995 |
| Q3 | Change Discrimination | Whether a meaningful change is present relative to the reference image. | 16,038 |
| Q4 | Change Localization | A regional or semantic location where the change occurs. | 8,991 |
| Q5 | Semantic Change Classification | The category of the observed change or defect. | 5,034 |
| Q6 | Change Description | The description that best summarizes the observed differences. | 8,995 |
| Q7 | Change Cause Inference | A plausible cause or underlying reason for the change. | 8,994 |

**Q1–Q2 assess single-image understanding; Q3–Q7 assess paired-image change analysis. All seven tasks use MCQ answer selection**, including Q2 and Q6. Q4 evaluates location-related answers, not pixel-level segmentation. Q7 evaluates plausible interpretation from available evidence, not experimentally established causality.

<div align="center">
  <img src="assets/cdbench-examples-v26.png" alt="Six CDBench examples covering image analysis and paired-image change analysis" width="100%">
  <p><em>Representative MCQs. Shaded options mark annotated answers. Panels (a)–(b) use single images; in (c)–(f), the reference/earlier image appears above the target/later image.</em></p>
</div>

## Dataset construction

1. **Integrate visual sources.** The benchmark combines seven datasets across three domains. Where a source lacks paired references, nearest-neighbor retrieval selects relevant normal examples. For remote-sensing images with limited prior information, CLIP-based scene classification supplies additional context.
2. **Draft and refine questions across models.** Available images, masks, labels, descriptions, and domain knowledge support question construction. GPT-4o drafts questions that Claude reviews and refines, and the models also exchange these roles.
3. **Verify with two experts.** Experts inspect and correct QA items for question relevance, answer validity, and ambiguity through a collaborative annotation platform.

This procedure is designed to reduce generator-specific bias and ambiguous answers. Masks and labels used to ground annotation are **not automatically inputs to the evaluated MLLMs**.

<div align="center">
  <img src="assets/cdbench-construction-v26.png" alt="Shared annotation context, reciprocal GPT-4o and Claude refinement, dual-expert verification, and seven MCQ tasks" width="100%">
  <p><em>CDBench construction and task organization. The verification cards illustrate a correct answer and an intentionally incorrect alternative; they are examples of the checks, not historical rejection records.</em></p>
</div>

## Repository files and availability

The current release contains `questions.csv`, six image archives, and documentation images in `assets/`:

```text
questions.csv
CDD.zip
LEVIR-CD.zip
SYSU-CD.zip
MVTec-AD.zip
MVTec-LOCO.zip
VisA.zip
assets/
README.md
```

**GoodsAD availability:** `questions.csv` includes 14,154 GoodsAD QA items covering 2,960 image pairs, but this repository does not currently contain a `GoodsAD.zip` archive. Downloading the six available archives therefore does not provide all images referenced by the CSV. For the matching GoodsAD files, contact the maintainers through the [Community tab](https://huggingface.co/datasets/MM-CD/CDBench/discussions). If evaluating only available subsets, report that subset explicitly.

The Hugging Face viewer exposes the annotations as the `viewer` configuration with a split named **`train`**. This is the current file-loading configuration; it does **not** define an official training/evaluation partition. The reported benchmark evaluates models without CDBench-specific fine-tuning.

### Annotation schema

| Field | Type | Meaning |
|---|---|---|
| `target_image` | string | Relative path to the target image. |
| `question_num` | integer | Task identifier, from 1 to 7. |
| `question` | string | Question stem in English. |
| `options` | string | Lettered answer choices, usually separated by newlines. |
| `answer` | string | Correct option letter: `A`, `B`, `C`, or `D`. |
| `reference_image` | string | Relative path to the reference image. |
| `mask` | string or empty | Relative path to an available change/defect mask; some entries are empty. |

Image fields contain **paths**, not embedded image bytes. The CSV contains quoted multiline option strings, so use a CSV parser rather than splitting the file by line. An empty mask field means no mask path is supplied for that row; it should not be treated as a negative-change label.

One released Q4 record is shown below. Newlines in `options` are escaped here for readability.

```json
{
  "target_image": "CDD/B/10_154.png",
  "question_num": 4,
  "question": "Where have the most significant changes occurred between Time Points A and B?",
  "options": "A. In the agricultural regions;\nB. Near large water bodies.;\nC. In the downtown urban area;\nD. Along the roadway infrastructure",
  "answer": "D",
  "reference_image": "CDD/A/10_154.png",
  "mask": "CDD/label/10_154.png"
}
```

## Quick start

### Load the annotations

```bash
pip install datasets huggingface_hub pillow
```

```python
from datasets import load_dataset

qa = load_dataset("MM-CD/CDBench", name="viewer", split="train")
print(qa.num_rows)  # 73122 in this release
print(qa[0]["question"])
print(qa[0]["options"])

# Example: select all regional-localization questions.
q4 = qa.filter(lambda row: row["question_num"] == 4)
```

Loading this CSV configuration retrieves the annotations; it does not download or decode the image archives.

### Download image archives

Download only the source datasets you need. For example:

```python
from huggingface_hub import hf_hub_download

archive = hf_hub_download(
    repo_id="MM-CD/CDBench",
    repo_type="dataset",
    filename="CDD.zip",
    local_dir="./CDBench",
)
print(archive)
```

Repeat with another filename from the repository listing as needed. Each archive is several gigabytes; the six archives total approximately 25.24 GB before extraction.

**Arrange extracted files to match the CSV paths.** Archive directory layouts differ: the remote-sensing archives start with an `mmcd_set/` wrapper, while the industrial archives start with object-category folders. Inspect the extracted tree and place each source under its dataset-name directory, avoiding extra wrapper levels. For example, the resulting image root should contain:

```text
images/
├── CDD/
│   ├── A/10_154.png
│   ├── B/10_154.png
│   └── label/10_154.png
├── LEVIR-CD/
├── SYSU-CD/
├── MVTec-AD/
├── MVTec-LOCO/
└── VisA/
```

After arranging the files, open a pair using the paths in its record:

```python
from pathlib import Path
from PIL import Image

image_root = Path("./CDBench/images")
row = qa[0]

with Image.open(image_root / row["reference_image"]) as image:
    reference = image.convert("RGB")
with Image.open(image_root / row["target_image"]) as image:
    target = image.convert("RGB")

print(reference.size, target.size)
```

The official [Hugging Face loading guide](https://huggingface.co/docs/datasets/loading) and [Hub download guide](https://huggingface.co/docs/huggingface_hub/guides/download) provide additional loading and revision-pinning options.

## Evaluation protocol

- **Metric:** top-1 MCQ accuracy for each of Q1–Q7. `Avg.` is the unweighted arithmetic mean of the seven task accuracies, not a pooled average over all QA rows.
- **Cascading scoring:** an incorrect foundational answer causes subsequent dependent answers for the same sample to be counted as incorrect. Later-task scores therefore reflect both the task itself and upstream judgments.
- **Setting:** zero-shot prompting without task-specific fine-tuning on CDBench. Each evaluation is run twice, and the reported scores average the two runs.
- **Inputs:** distinguish the single-image tasks from paired-image tasks. Keep answers and annotation-only information out of model prompts; report any additional masks, retrieved context, or expert guidance used by a system.

For comparable results, report the evaluated subsets, model version, prompts, input context, per-task scores, and scoring protocol. Independent per-question accuracy without cascading is a different metric. The current repository provides data files; it does not contain a released scoring script or machine-readable task-dependency specification.

### Reported model performance

Accuracy (%). **Bold marks the best value in each column.** These are the reported benchmark runs, not evaluations of the latest versions of the named model families.

| Model | Q1 | Q2 | Q3 | Q4 | Q5 | Q6 | Q7 | Avg. |
|---|---:|---:|---:|---:|---:|---:|---:|---:|
| Qwen-Max-Latest | 57.03 | 79.97 | 55.88 | 72.33 | 61.78 | 64.35 | 65.00 | 65.19 |
| Qwen-plus-latest | 55.12 | 79.04 | 55.87 | 72.18 | 62.77 | 63.70 | 65.93 | 64.94 |
| Qwen-turbo-1101 | 48.90 | 68.15 | 55.93 | 66.07 | 54.63 | 57.25 | 59.82 | 58.68 |
| Grok-3 | 73.16 | 80.38 | 59.67 | 68.51 | 60.58 | 61.70 | 62.26 | 66.61 |
| Claude-3-5-sonnet | 70.17 | 82.27 | 67.47 | 71.74 | 64.96 | 58.48 | 64.15 | 68.46 |
| Claude-3-7-sonnet | 65.87 | 75.26 | **79.87** | 65.40 | 61.58 | 59.92 | 61.26 | 67.02 |
| Gemini-1.5-flash | 48.21 | 70.15 | 57.30 | 63.51 | 60.00 | 55.81 | 56.15 | 58.73 |
| Gemini-1.5-pro | 72.54 | 83.07 | 66.03 | 73.96 | 66.55 | 67.59 | 68.60 | 71.19 |
| GPT-4o | 91.37 | **85.00** | 69.72 | 69.70 | 56.83 | 61.71 | 60.46 | 70.68 |
| ChangeAgent | **96.87** | 76.78 | 78.81 | **76.79** | **77.67** | **70.82** | **69.99** | **78.25** |

ChangeAgent improves the average by **7.06 percentage points** over Gemini 1.5 Pro and **7.57 points** over GPT-4o. It ranks first on Q1 and Q4–Q7; GPT-4o leads Q2, and Claude-3-7-sonnet leads Q3.

<div align="center">
  <img src="assets/cdbench-performance-v26.png" alt="Model comparison using the same ten model rows and accuracies as the table" width="100%">
  <p><em>Task-group means, overall means, and individual Q1–Q7 accuracies from the table above. Shading identifies ChangeAgent.</em></p>
</div>

### ChangeAgent

ChangeAgent combines paired CLIP visual features, an aggregated change map, retrieved domain knowledge, and task inputs for MLLM reasoning. The change map supplies spatial evidence, while the original visual features retain broader scene semantics. Retrieval adds relevant priors, task definitions, and change examples.

<div align="center">
  <img src="assets/changeagent-framework-v26.png" alt="ChangeAgent architecture with shared visual encoding, change decoding, retrieval augmentation, and MLLM reasoning" width="100%">
  <p><em>ChangeAgent uses both visual guidance and retrieved knowledge for seven-task reasoning.</em></p>
</div>

Because it uses additional visual guidance and retrieval, ChangeAgent serves as a **hybrid reference**, rather than an input-matched MLLM baseline. The reported comparison does not isolate the contribution of each component; that requires ablation experiments.

### Visual dependence and language bias

GPT-4o was evaluated with and without images on the **same randomly sampled 1,000 image pairs (4,628 QA items)** under the cascading protocol.

| Input | Avg. (Q1–Q7) | Q4 | Q5 | Q6 | Q7 |
|---|---:|---:|---:|---:|---:|
| With images | 70.68 | 69.70 | 56.83 | 61.71 | 60.46 |
| Text-only | 32.93 | 26.39 | 15.74 | 22.74 | 21.88 |

Removing images reduces the average by **37.75 percentage points**, showing a substantial contribution from visual evidence. This experiment does **not** establish the absence of residual language bias. Cascading also propagates upstream errors, so later-task scores should be interpreted in that context.

## Scope and limitations

CDBench measures structured visual comparison and answer selection across the included domains. It complements pixel-level change-detection benchmarks and free-form captioning evaluations; its accuracy should not be interpreted as segmentation quality or unrestricted generation quality. Retrieved normal references in anomaly datasets are not necessarily temporal observations of the same object, and the image pairs are not universally registered.

The source datasets, task frequencies, and answer formats are heterogeneous. Report domain and task coverage when comparing systems. Cross-model refinement and expert verification aim to improve annotation quality, but do not guarantee that all ambiguity, language priors, or model-specific bias has been eliminated.

## Project and demo

- [Project website](http://mm-cd.org)
- [Interactive demo](http://demo.mm-cd.org:8880) — availability and response times depend on the demo server.
- [Community discussions](https://huggingface.co/datasets/MM-CD/CDBench/discussions) for data issues, missing files, and questions.

## License and acknowledgements

This repository is distributed with the **CC BY-NC-SA 4.0** license tag. Images originate from the seven source datasets listed above; consult their original licenses and attribution requirements as well.

We thank the creators of LEVIR-CD, SYSU-CD, CDD, MVTec-AD, MVTec-LOCO, VisA, and GoodsAD; the experts who annotated and verified the QA items; Fudan University; Shanghai Ocean University; Shanghai Vision Medical Technology Co., Ltd.; and the open-source community supporting the underlying vision-language models and tools.

## Citation

To cite this dataset release:

```bibtex
@misc{chen2026cdbench,
  title  = {{CDBench}: A Comprehensive Multimodal Dataset and Evaluation Benchmark for General Change Detection},
  author = {Chen, Jintao and Qian, Rui and Peng, Bo and Cheng, Linjie and Li, Xiang and Chen, Tian and Ye, Jingyong and Yang, Yinhao and Chi, Mingmin and Dou, Dejing},
  year   = {2026},
  url    = {https://huggingface.co/datasets/MM-CD/CDBench},
  note   = {Dataset and benchmark}
}
```