Train two-wheeler detection and road-perception models with annotations focused on electric and conventional bicycles.

Train models to generate clearly structured advertising text for outdoor media using verifiable layout constraints across multiple display formats.

The cabin voice assistant control command dataset is designed specifically for optimizing voice recognition and control in vehicular driving.

Train models to reason over highway lane topology and accurately determine whether lanes and exits are connected through multi-hop question answering.

The Adverse Weather Road Image Classification Dataset focuses on enhancing the performance of intelligent transportation systems under complex meteorological conditions.

This dataset is specifically designed for autonomous driving and intelligent transportation systems, containing high-precision annotated lane line image data.

Trace out-of-home ad claims from claim extraction through evidence linking to an overall fact-review conclusion.

Train vehicle recognition models and support transport safety monitoring with vehicle images, object annotations, and class information.

Train models to distinguish current collision damage from pre-existing damage by integrating accident, claim, and inspection evidence.

Explore end-to-end tool-use traces for LNG carrier port-call coordination and train agents to gather, validate, and integrate information from multiple sources.

This dataset contains high-definition videos for analyzing and optimizing watering operations.

Build smarter railway inspections with track surface images and defect annotations designed to support automated recognition.

Paired original topology, error analysis, and repair targets teach models to identify intersection issues and generate consistent topology structures.

The School Area Noise Pollution Assessment Dataset focuses on analyzing and evaluating campus noise levels.

Evaluate how well models understand cross-layer positions, alignment, and containment in HD maps featuring lane markings and road boundaries.

Turn road signs, lane restrictions, vehicle classes, and time conditions into queryable rule relations for Robotaxi maps.

Evaluate how well models understand road sign locations along 2D routes using travel direction, sign order, and left-right annotations.

Test whether models can extract key passenger railcar specifications from product pages and link them to the correct model version.

Develop in-cabin safety perception with driver images and annotation metadata for behavior recognition and state monitoring.

Turn the spatial relationship between a vehicle and its lane centerline into clear, trainable lane-relative localization fields.