Datasets › VideoDB's OCR Benchmark Public Collection

VideoDB's OCR Benchmark Public Collection

Introduced by Sankalp Nagaonkar et al. in Benchmarking Vision-Language Models on Optical Character Recognition in Dynamic Video Environments10 Feb 2025 archive 2025-07-28

Dataset Introduction

This dataset leverages VideoDB's Public Collection to offer a diverse range of videos featuring text-containing scenes. It spans multiple categories—ranging from finance and legal documents to software UI elements and handwritten notes—ensuring a broad representation of real-world text appearances. Each video is annotated with frame indexes to facilitate consistent and reproducible OCR benchmarks. Currently, the dataset includes over 25 curated videos, yielding thousands of extracted frames that present a variety of text-related challenges.

Key Features
  1. Diverse Text Genres

    • Finance/Business: Includes news tickers and stock market visuals where text scrolls rapidly.
    • Legal/Educational: Features documents with formal language, diagrams, and formatted text.
    • Software/Web Development/UI: Shows on-screen code editors, browser windows, and other UI elements that test OCR's ability to handle varying font sizes and code snippets.
    • Handwriting: Encompasses both cursive and print handwriting on whiteboards or paper, capturing the challenges of style variability and penmanship.
    • Miscellaneous/Other: Covers signage, billboards, and everyday text in the wild.
  2. Rich Annotation

    • Each video includes frame indexes or scene timestamps to ensure consistent, reproducible extraction of text segments.
    • Ground truth (OCR text) is provided for thousands of extracted frames, facilitating quantitative performance evaluations (e.g., CER, WER).
  3. Benchmark-Ready

    • The dataset seamlessly integrates with the ocr-benchmark repository to streamline model evaluation.
    • Scripts are included for frame extraction, automatic OCR comparison, and metric calculation (CER, WER, accuracy).
How to Access
  • VideoDB Public Collection ID: c-c0a2c223-e377-4625-94bf-910501c2a31c
    Simply reference this ID within VideoDB to retrieve and review the videos.
  • Ground Truth Files: Located in the ocr_ground_truths directory. The JSON files map each frame to its corresponding textual annotations.

For detailed instructions on working with VideoDB Public Collections, please refer to the official documentation.

Licensing and Usage

Benchmarks archive 2025-07-28

All 1 leaderboard whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

First row (archive order)PaperCode
Optical Character Recognition (OCR) VideoDB's OCR Benchmark Public Collection GPT-4o Average Accuracy 76.22 Benchmarking Vision-Language Models on Optical Character... video-db/ocr-benchmark 5 Compare

Papers archive 2025-07-28

papers with a benchmark row: 1 · with a code link: 1 · where Syntology ran a sample: 1 (1 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument) Syntology

Show: all papers with a benchmark rowonly where code ran (1 of 1 with a benchmark row: 1 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument)

Page 1 of 1: paper 1 of 1 with a leaderboard row on this dataset's benchmarks, newest first by the archive's date (ties by slug; undated papers last).

The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset, not that list; the archive's count for this dataset is 1. The Syntology column is from Syntology's graph, stated per sample; it is not part of any archive number. A line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified” (C is a part of N, never taken away from it); the figure “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced. When the archive marks a repository official for the paper, the cell starts with that repository's state (the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record.

Dataset loaders archive 2025-07-28

No loader listed in the archive.

Tasks archive 2025-07-28

License archive 2025-07-28

No licence recorded in the archive. Absence here is not a statement about the dataset's terms.

Modalities archive 2025-07-28

Languages archive 2025-07-28

Variants archive 2025-07-28

  • VideoDB's OCR Benchmark Public Collection

1 variant name, as the archive lists them.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections