Sign in to view source links and access this dataset
Description
45,986 high-resolution images with approximately 481,947 bounding box annotations for 14 vehicle classes, captured from operational CCTV cameras in Bengaluru, India. BMD-45 is a large-scale, India-specific benchmark for urban traffic analysis created by iisc-aim and last updated in April 2026.
Use Cases
Train object detection models based on the 14 fine-grained vehicle categories.
Benchmark model performance on Indian urban traffic scenes using the large-scale annotated images.
Analyze traffic composition and density patterns in a highly congested megacity based on the CCTV camera data.
Strengths
Large scale with 45,986 high-resolution (1920×1080) images.
Substantial annotation volume with approximately 481,947 bounding boxes.
Captures real-world diversity from 3,679 operational CCTV cameras across a major city.
Focus on a specific, high-traffic geography (Bengaluru, India).
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Data may reflect geographic and temporal bias inherent to the specific CCTV camera network in Bengaluru.
Provenance
Source
iisc-aim
Collection Method
Captured from operational CCTV cameras across Bengaluru.
Freshness
Last updated 2026-04-10 18:05:26; freshness should be verified.
Geography
Bengaluru, India
License is unknown and should be verified before use.