Datasets:

oValach
/

Crash-VQA

Tasks:

Visual Question Answering

Modalities:

Image

Text

Tabular

Sub-tasks:

visual-question-answering

closed-domain-qa

multi-class-classification

Languages:

Size:

Tags:

Libraries:

License:

Dataset card Data Studio Files Files and versions

xet

Community

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Crash-VQA access is gated to document noncommercial research use and attribution/citation agreement.

Crash-VQA

Crash-VQA: A Multi-View Dataset for Post-Crash Mechanics and Severity Prediction

Crash-VQA multi-view vehicle example

Nine canonical views of one damaged vehicle. Missing views are represented as null.

Overview

Crash-VQA is a multi-view benchmark for structured prediction of post-crash mechanics and severity. Each task instance contains one prediction target, a normalized label, optional structured vehicle metadata, and up to nine canonical views of the same vehicle. The natural configuration defines the complete benchmark, while balanced is a derived configuration for controlled task and label balancing.

At a glance

Metric	Value
Task instances	48,871
Unique images across the release	97,323
Unique cases	9,913
Unique vehicles	12,734
Tasks	5
Canonical image views	9

The natural configuration defines the complete Crash-VQA benchmark and contains 48,871 task instances. The balanced configuration is a derived view of the corresponding train and test partitions for controlled task and label balancing. It overlaps with natural and therefore does not increase the number of unique benchmark instances.

What makes Crash-VQA useful

Multi-view evidence: up to nine canonical views provide complementary observations of the same damaged vehicle.
Structured targets: all five tasks use compact normalized label spaces suitable for reproducible evaluation.
Benchmark and derived view: natural defines the complete benchmark, while balanced is an overlapping derived configuration for controlled task and label comparisons.
Vehicle context: optional vehicle type, model year, and curb weight are available as structured metadata.
Vehicle-based splits: all task instances associated with the same vehicle remain in the same split.

Example task rows

Representative Crash-VQA task examples

The montage shows one representative image for each task. A model receives the available multi-view image set and predicts the task-specific label.

Quick start

from datasets import load_dataset

natural = load_dataset("oValach/Crash-VQA", "natural")
balanced = load_dataset("oValach/Crash-VQA", "balanced")

row = natural["train"][0]
print(row["task"], row["label"], row["metadata"])

available_images = {
    name: row[name]
    for name in ['image_front', 'image_rear', 'image_right', 'image_left', 'image_front_right', 'image_front_left', 'image_rear_right', 'image_rear_left', 'image_top']
    if row[name] is not None
}

Explore the dataset

Tasks

Task	Prediction target	Allowed labels	Task instances
`plane_atomic`	Principal impact plane	`front`, `rear`, `left`, `right`	12,563
`clock_atomic`	Impact direction on a clock face	`1 through 12`	12,459
`extent_atomic`	Damage extent	`minor`, `moderate`, `severe`	11,455
`deltav_atomic`	Delta-V interval	`0-10`, `10-20`, `20-30`, `30+`	7,745
`ais2_atomic`	AIS 2+ injury indicator	`true`, `false`	4,649

Configurations and splits

Configuration	Split	Rows
`natural`	`train`	36,632
`natural`	`validation`	4,902
`natural`	`test`	7,337
`balanced`	`train`	13,202
`balanced`	`test`	2,414

natural is the default benchmark configuration and contains 48,871 task instances across train, validation, and test splits.
balanced is a derived configuration with 13,202 training rows and 2,414 test rows. It overlaps with natural and is intended for controlled task and label comparisons.

Dataset structure

Each materialized row represents one task instance.

Column	Type	Description
`case_id`	string	Stable public case identifier
`split`	string	Train, validation, or test
`task`	string	One of the five task names
`label`	string	Normalized task label
`metadata`	struct	Vehicle type, model year, and curb weight
`image_front`	`Image()` or null	Front vehicle view
`image_rear`	`Image()` or null	Rear vehicle view
`image_right`	`Image()` or null	Right vehicle view
`image_left`	`Image()` or null	Left vehicle view
`image_front_right`	`Image()` or null	Front-right vehicle view
`image_front_left`	`Image()` or null	Front-left vehicle view
`image_rear_right`	`Image()` or null	Rear-right vehicle view
`image_rear_left`	`Image()` or null	Rear-left vehicle view
`image_top`	`Image()` or null	Top vehicle view
`vehicle_id`	string	Stable public vehicle identifier
`sample_id`	string	Stable task-instance identifier

The nine image columns are Hugging Face Image() features. Path-backed cells use repository-relative paths; the raw Parquet representation may contain a nullable bytes member, while the Dataset Viewer renders the image itself.

Image views

Column	Canonical view
`image_front`	Front
`image_rear`	Rear
`image_right`	Right
`image_left`	Left
`image_front_right`	Front-right
`image_front_left`	Front-left
`image_rear_right`	Rear-right
`image_rear_left`	Rear-left
`image_top`	Top

Vehicle metadata

Field	Type	Meaning
`vehicle_type`	string or null	Source vehicle body/type description
`model_year`	int64 or null	Vehicle model year
`curb_wt_kg`	float64 or null	Curb weight in kilograms

Nulls indicate unavailable source values.

Source and provenance

The source imagery and vehicle/crash metadata originate from the National Highway Traffic Safety Administration Crash Investigation Sampling System (CISS). CISS collects detailed information from a representative sample of crashes to support vehicle-safety research. Crash-VQA adds the public task formulation, normalized labels, split definitions, identifiers, packaging, and documentation.

See SOURCE_NOTICE.md for source attribution, source terms, and the non-endorsement notice.

Intended uses

Crash-VQA is intended for:

multi-view, instruction-conditioned visual prediction research;
post-crash mechanics and vehicle damage understanding;
structured post-crash mechanics and severity benchmarking;
evaluation of multimodal models across image and vehicle-metadata inputs;
controlled comparison using the natural benchmark and derived balanced configuration.

Out-of-scope uses

Crash-VQA should not be used as the sole basis for:

medical, legal, insurance, or accident-liability decisions;
identifying people, owners, or specific crash locations;
operational vehicle-safety certification;
claims that exceed the observable evidence or supplied metadata.

Limitations and responsible use

The dataset may reflect sampling, reporting, geography, vehicle-fleet, crash-severity, image-quality, and missing-metadata biases from the source collection. Damage appearance may be ambiguous across views, and labels simplify complex crash phenomena into discrete benchmark targets. Users should report performance separately by task and should not interpret benchmark performance as evidence of real-world crash-reconstruction capability.

License

Crash-VQA contains two licensing and provenance layers:

Underlying NHTSA/CISS source material: source imagery and vehicle/crash metadata originate from NHTSA CISS. These source materials are not relicensed by Crash-VQA and remain subject to applicable NHTSA/DOT terms, disclaimers, and source notices. See SOURCE_NOTICE.md.
Crash-VQA original contributions: original annotations, normalized labels, task definitions, split definitions, public identifiers, organization, packaging, dataset card content, and documentation are licensed under CC BY-NC 4.0, to the extent those rights are controlled by the Crash-VQA authors. See LICENSE.

Copies validly obtained under the earlier CC BY 4.0 release remain governed by CC BY 4.0. This license change is not retroactive.

Paper

The accompanying paper, “Crash-VQA: A Multi-View Dataset for Post-Crash Mechanics and Severity Prediction,” has been accepted at the MARS² Workshop at ECCV 2026.

Citation

The official publication citation will be added once the workshop proceedings are publicly available. Until then, please cite this dataset repository and refer to the accompanying paper by the title above.

Downloads last month: 120

Total file size:

50.4 GB

Models trained or fine-tuned on oValach/Crash-VQA

oValach/CrashQwen-8B

Image-Text-to-Text • Updated 6 days ago