Overview
ISO/IEC 15938-3:2002/Amd 4:2010 - part of the MPEG‑7 family - amends the Visual part of the Multimedia Content Description Interface to add video signature tools. It extends the visual descriptor set (color, texture, shape, motion, localization) and the basic descriptor containers (GridLayout, TimeSeries, GofGopFeature, MultipleView) and supporting tools (Spatial2DcoordinateSystem, TemporalInterpolation). The amendment defines a VideoSignature descriptor designed for robust, unique identification of video content (video fingerprinting) and specifies binary codec classification entries for many visual descriptors.
Key topics and technical requirements
- Video signature architecture
- Frame signature: a 380‑dimensional vector of base‑3 (ternary) values representing intensity relations and statistical operators (mean / first / second order).
- Words: compact 1‑byte summaries derived from frame signatures; vocabulary of possible words.
- Bag‑of‑words: group‑of‑frames representation capturing word occurrences over temporal segments to provide temporal robustness (order‑insensitive).
- Support for single or multiple static spatial regions per frame.
- Visual descriptor families
- Color: DominantColor, ScalableColor, ColorLayout, ColorStructure, ColorTemperature; supporting tools like ColorSpace, ColorQuantization, IlluminationInvariantColor, GoFGoPColor.
- Texture: HomogeneousTexture, TextureBrowsing, EdgeHistogram.
- Shape: RegionShape, ContourShape, ShapeVariation (temporal), Shape3D, Perceptual3DShape.
- Motion: CameraMotion, MotionTrajectory, ParametricMotion, MotionActivity.
- Localization: RegionLocator, SpatioTemporalLocator.
- Face descriptors: FaceRecognition, AdvancedFaceRecognition.
- Signatures and binary codecs
- ImageSignature and VideoSignature descriptors for robust identification.
- Classification scheme entries (MPEG7* codec names) and assignment of descriptor IDs (e.g., VideoSignature ID = 34) for binary interoperability.
- Interoperability
- Use of the codec configuration mechanism defined in ISO/IEC 15938‑1 for signaling video‑specific binary syntax.
Applications and practical value
- Duplicate detection / video fingerprinting for copyright enforcement, content takedown, and automated moderation.
- Large‑scale video search and retrieval in media asset management (MAM), streaming platforms, and search engines using content‑based indexing.
- Forensics and provenance tracking by identifying re‑encoded or edited copies of the same source.
- CDN deduplication and storage optimization by recognizing identical or near‑identical content.
- Surveillance and archival systems for fast lookup of similar video segments.
Who should use this standard
- Multimedia software developers and engineers implementing content‑based indexing or fingerprinting.
- Metadata architects and MAM/CDN vendors integrating MPEG‑7 descriptors.
- Legal/copyright teams and content moderation platforms needing reliable duplicate detection.
- Research groups working on video retrieval, computer vision, and multimedia forensics.
Related standards
- ISO/IEC 15938 (MPEG‑7 family) - overall multimedia content description.
- ISO/IEC 15938‑1 - Systems / codec configuration mechanism referenced for binary signaling.
Keywords: ISO/IEC 15938-3, MPEG-7, video signature, video fingerprinting, visual descriptors, image signature, multimedia content description, color descriptors, texture descriptors, motion descriptors.