UniOcc: A Unified Benchmark for Occupancy Forecasting and Prediction in Autonomous Driving

ICCV 25 2025 51.6 method

TLDR

UniOcc is a unified benchmark for occupancy forecasting and prediction in autonomous driving, integrating real-world and simulated data with novel metrics.

Reasoning

The paper's strength lies in its comprehensive unification of multiple real-world and simulated datasets with novel evaluation metrics. However, the abstract lacks specific quantitative results and does not discuss limitations.

Read-first score

Read-first score 51.6, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 1.

Methodology quality 25%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=benchmark,dataset,evaluation,experiment,metric

Recency 8%

86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=True; dataset=False; markers=code,dataset,github

Topical relevance 42%

1.4

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Field roles

FrontierMethodology anchorReproducibility anchor

Rank sensitivity

Stability: volatile; rank range: 872.

Keyword Scores

world dynamics prediction

world model

world simulator

generative world model

interactive world model

video world model

model-based reinforcement learning world model

Deep Analysis

Innovations

Unified benchmark and toolkit combining multiple real-world datasets (nuScenes, Waymo) and high-fidelity driving simulators (CARLA, OpenCOOD) for both occupancy forecasting and prediction.
Novel evaluation metrics that do not depend on ground-truth labels, enabling robust assessment of occupancy quality.
Annotation of per-voxel flows in occupancy labels, providing explicit flow information.
Comprehensive support for two tasks: occupancy forecasting (predicting future occupancies from historical data) and occupancy prediction (predicting current occupancy from camera images).

Methodology

UniOcc unifies data from nuScenes, Waymo, CARLA, and OpenCOOD, providing 2D/3D occupancy labels and per-voxel flows. It introduces novel evaluation metrics independent of ground-truth labels. Experiments are conducted on state-of-the-art models to evaluate the impact of large-scale diverse training data and explicit flow information.

Key Results

Large-scale, diverse training data and explicit flow information significantly enhance occupancy prediction and forecasting performance.

Resources

Paper Code