| --- |
| pretty_name: "Verified Object Counts" |
| license: other |
| license_name: kled-commercial-training-license |
| license_link: https://www.kled.ai/datasets/cluttered-surfaces-for-counting |
| task_categories: |
| - object-detection |
| - visual-question-answering |
| tags: |
| - kled |
| - commercial-license |
| - consented-data |
| - human-collected |
| - sample-on-request |
| - image |
| - real-world-capture |
| - phone-capture |
| - global |
| - labeled |
| - robotics |
| - manipulation |
| - counting |
| - vlm-evaluation |
| size_categories: |
| - 10K<n<100K |
| annotations_creators: |
| - crowdsourced |
| language_creators: |
| - crowdsourced |
| source_datasets: |
| - original |
| viewer: false |
| gated: manual |
| extra_gated_heading: "Request a sample set of Verified Object Counts" |
| extra_gated_prompt: "Kled delivers sample sets and licensed data by email, not through Hugging Face. Tell us a little about your work and the Kled data team will send a sample set and pricing. Your request goes to team@kled.ai." |
| extra_gated_fields: |
| Company or lab: text |
| Role: text |
| What are you building: text |
| Volume you need: |
| type: select |
| options: |
| - Sample set only |
| - Under 10,000 items |
| - 10,000 to 100,000 items |
| - 100,000 or more items |
| - Full dataset |
| - Custom collection on the Kled contributor base |
| I agree to be contacted by Kled about this request: checkbox |
| extra_gated_button_content: "Request a sample set" |
| extra_gated_description: "Prefer email? Write to team@kled.ai with the dataset name in the subject line." |
| --- |
| # Dataset Card for Verified Object Counts |
|
|
| <p align="center"><img src="https://knrucgucrxvizxlraxgn.supabase.co/storage/v1/object/public/task-images/v3/54/header-fb2c7593.jpg" alt="Verified Object Counts" width="520"></p> |
|
|
| > Overhead photos of cluttered surfaces with ten or more objects, each with an exact count entered by the photographer. |
|
|
| Collected through the Kled app by consenting contributors and licensed for commercial training. The files are delivered by Kled under license, not downloaded from this repository. Use **Request a sample set** at the top of this page, or write to [team@kled.ai](mailto:team@kled.ai). Catalog entry: [kled.ai/datasets/cluttered-surfaces-for-counting](https://www.kled.ai/datasets/cluttered-surfaces-for-counting?utm_source=huggingface&utm_medium=dataset_card&utm_campaign=cluttered-surfaces-for-counting). |
|
|
| ## Dataset Details |
|
|
| ### Dataset Description |
|
|
| Find a surface with at least ten countable objects, photograph it from above, count carefully, enter the exact number. The count is checked on review, and images with fewer than ten objects, heavy occlusion or a wrong count are rejected; ten photos per contributor at most, worldwide, so the class mix is wide. The count is the only label; boxes or dots can be added cheaply because the count tells annotators when they are done. Counting is a perceptual task VLMs still get wrong: a December 2025 evaluation on FSC-147 (arXiv 2512.15254) put the best model near 0.5 exact-count accuracy, and FSC-147 itself is 6,000 images under unstated terms. |
|
|
| - **Curated by:** Kled ([useKled](https://huggingface.co/useKled)) |
| - **Collected through:** Phone camera (in app) |
| - **Regions:** Global, 100+ countries |
| - **Volume:** 25,000+ images (approx. 75 GB) |
| - **Contributors:** 5,000+ |
| - **Modality:** Images |
| - **Format:** JPG / PNG / HEIC |
| - **Resolution:** Native phone camera resolution (JPG, PNG, HEIC) |
| - **Annotations:** Labeled (human-verified object count) |
| - **Quality control:** 99% passed QA review |
| - **License:** Non-exclusive commercial training license ([terms](https://www.kled.ai/datasets/cluttered-surfaces-for-counting?utm_source=huggingface&utm_medium=dataset_card&utm_campaign=cluttered-surfaces-for-counting)); exclusivity available on request |
| - **Consent:** Each contributor opted in to this use when they uploaded. |
|
|
| ### Dataset Sources |
|
|
| - **Catalog page:** [kled.ai/datasets/cluttered-surfaces-for-counting](https://www.kled.ai/datasets/cluttered-surfaces-for-counting?utm_source=huggingface&utm_medium=dataset_card&utm_campaign=cluttered-surfaces-for-counting) |
| - **Sample set:** through the access request at the top of this page, or [team@kled.ai](mailto:team@kled.ai) |
| - **Publisher:** [huggingface.co/useKled](https://huggingface.co/useKled) |
|
|
| ## Uses |
|
|
| ### Direct Use |
|
|
| - Object counting and class-agnostic counting evaluation |
| - Dense detection in clutter |
| - VLM numeracy benchmarks |
|
|
| ### Out-of-Scope Use |
|
|
| Redistribution or resale of the files, use outside the terms of the commercial training license, and any attempt to identify or contact contributors or people who appear in the data. Uses that require labels this set does not carry should be scoped with the data team first. |
|
|
| ## Dataset Structure |
|
|
| ### Data Instances |
|
|
| One image per record, delivered as JPG / PNG / HEIC. Volume: 25,000+ images. |
|
|
| ### Data Fields |
|
|
| - **File:** the JPG / PNG / HEIC file as captured or uploaded by the contributor |
| - **Labels:** Labeled (human-verified object count), entered by the contributor at capture and checked at review. |
| - **Capture metadata:** capture tool, country of capture, submission time |
| - **Review status:** QA outcome (99% passed QA review) |
|
|
| Field names and the manifest format are confirmed with the sample set. |
|
|
| ### Data Splits |
|
|
| Delivered as one set. Subsets by country, language or label, and custom splits, are available on request. |
|
|
| ## Dataset Creation |
|
|
| ### Curation Rationale |
|
|
| The dataset exists because the public alternatives named in the description are non-commercial, synthetic, small or collected elsewhere. Kled runs the task as a paid brief in its app so that the data is captured on purpose, under the capture rules above, by people who have opted in to commercial use. |
|
|
| ### Source Data |
|
|
| #### Data Collection and Processing |
|
|
| Contributors complete the task in the Kled app. Capture tool: Phone camera (in app). Each submission is reviewed against the task brief; submissions that break the capture rules are rejected. 99% passed QA review. Files are delivered as JPG / PNG / HEIC, in the resolution noted above. |
|
|
| #### Who are the source data producers? |
|
|
| Kled contributors, who complete paid data tasks on their own phones. This task was open to contributors worldwide. Kled has 625,000+ contributors in 100+ countries, strongest in Southeast Asia, the United States and Europe. |
|
|
| ### Annotations |
|
|
| #### Annotation process |
|
|
| Labeled (human-verified object count). The label is entered by the contributor as part of the task and checked at review against the task brief; it is a contributor label, not an expert annotation, unless the description above says otherwise. |
|
|
| #### Who are the annotators? |
|
|
| The contributors themselves, at the moment of capture. Kled reviewers check each submission against the task brief. |
|
|
| #### Personal and Sensitive Information |
|
|
| Every item was contributed by its owner, who opted in to this use at upload. The task brief states what may and may not appear in a submission, and reviewers reject submissions that break those rules. Kled's data handling practices are documented at [trust.kled.ai](https://trust.kled.ai/); questions about handling requirements for your use go to [team@kled.ai](mailto:team@kled.ai). |
|
|
| ## Bias, Risks, and Limitations |
|
|
| Coverage follows the contributor base: strongest in Southeast Asia, the United States and Europe, thinner elsewhere, and not limited by region for this task. Devices, settings and conditions vary by contributor, which is the point of the collection and also means no two items are captured under identical conditions. Where labels are contributor-entered, they are checked at review but are not expert annotations. Volumes are the current scope of the task and grow as collection continues. |
|
|
| ### Recommendations |
|
|
| Request a sample set and run it against your task before licensing. Ask for per-country or per-language counts where they matter to your use. |
|
|
| ## Licensing Information |
|
|
| Non-exclusive commercial training license; exclusivity on request. Each contributor opted in to this use when they uploaded. The license terms are on the [catalog page](https://www.kled.ai/datasets/cluttered-surfaces-for-counting?utm_source=huggingface&utm_medium=dataset_card&utm_campaign=cluttered-surfaces-for-counting). |
|
|
| ## Citation |
|
|
| ```bibtex |
| @misc{kled_cluttered_surfaces_for_counting_2026, |
| title = {Verified Object Counts}, |
| author = {Kled}, |
| year = {2026}, |
| url = {https://www.kled.ai/datasets/cluttered-surfaces-for-counting}, |
| note = {Commercially licensed dataset collected through the Kled app. Sample sets on request.} |
| } |
| ``` |
|
|
| ## Dataset Card Contact |
|
|
| Kled data team: [team@kled.ai](mailto:team@kled.ai) |
|
|
| ## Related Kled datasets |
|
|
| - [Counterfeit and Genuine Pairs (5,000 pairs)](https://huggingface.co/datasets/useKled/counterfeit-vs-authentic-products) |
| - [Near-Identical Object Groups (5,000 images)](https://huggingface.co/datasets/useKled/fine-grained-similar-objects) |
| - [Object Removal Sequences (10,000 sequences)](https://huggingface.co/datasets/useKled/take-one-thing-away-then-shoot-again) |
| - [Eight-View Object Orbits (10,000 objects)](https://huggingface.co/datasets/useKled/walk-around-one-object) |
|
|
