51554947a0
Video Schema Classes (9 files): - VideoPost, VideoComment: Social media video modeling - VideoTextContent: Base class for text content extraction - VideoTranscript, VideoSubtitle: Text with timing and formatting - VideoTimeSegment: Time code handling with ISO 8601 duration - VideoAnnotation: Base annotation with W3C Web Annotation alignment - VideoAnnotationTypes: Scene, Object, OCR detection annotations - VideoChapter, VideoChapterList: Navigation and chapter structure - VideoAudioAnnotation: Speaker diarization, music, sound events Enumerations (12 enums): - VideoDefinitionEnum, LiveBroadcastStatusEnum - TranscriptFormatEnum, SubtitleFormatEnum, SubtitlePositionEnum - AnnotationTypeEnum, AnnotationMotivationEnum - DetectionLevelEnum, SceneTypeEnum, TransitionTypeEnum, TextTypeEnum - ChapterSourceEnum, AudioEventTypeEnum, SoundEventTypeEnum, MusicTypeEnum Examples (904 lines, 10 comprehensive heritage-themed examples): - Rijksmuseum virtual tour chapters (5 chapters with heritage entity refs) - Operation Night Watch documentary chapters (5 chapters) - VideoAudioAnnotation: curator interview, exhibition promo, museum lecture All examples reference real heritage entities with Wikidata IDs: Q5598 (Rembrandt), Q41264 (Vermeer), Q219831 (The Night Watch)
375 lines
12 KiB
YAML
375 lines
12 KiB
YAML
# Video Time Segment Class
|
|
# Reusable temporal segment for video content (subtitles, annotations, chapters)
|
|
#
|
|
# Part of Heritage Custodian Ontology v0.9.5
|
|
#
|
|
# STRUCTURE:
|
|
# VideoTimeSegment (this class)
|
|
# - start_time, end_time (ISO 8601 duration)
|
|
# - start_seconds, end_seconds (float for computation)
|
|
# - segment_text (text content for this segment)
|
|
# - confidence (for ASR/CV generated content)
|
|
#
|
|
# USED BY:
|
|
# - VideoSubtitle (time-coded caption entries)
|
|
# - VideoAnnotation (scene/object detection segments)
|
|
# - VideoChapter (user-defined chapters)
|
|
#
|
|
# ONTOLOGY ALIGNMENT:
|
|
# - Maps to Media Fragments URI 1.0 (W3C) for temporal addressing
|
|
# - CIDOC-CRM E52_Time-Span for temporal extent
|
|
# - Web Annotation oa:FragmentSelector for annotation targets
|
|
|
|
id: https://nde.nl/ontology/hc/class/VideoTimeSegment
|
|
name: video_time_segment_class
|
|
title: Video Time Segment Class
|
|
|
|
imports:
|
|
- linkml:types
|
|
|
|
prefixes:
|
|
linkml: https://w3id.org/linkml/
|
|
hc: https://nde.nl/ontology/hc/
|
|
schema: http://schema.org/
|
|
dcterms: http://purl.org/dc/terms/
|
|
crm: http://www.cidoc-crm.org/cidoc-crm/
|
|
oa: http://www.w3.org/ns/oa#
|
|
ma: http://www.w3.org/ns/ma-ont#
|
|
|
|
default_prefix: hc
|
|
|
|
classes:
|
|
|
|
VideoTimeSegment:
|
|
class_uri: crm:E52_Time-Span
|
|
abstract: false
|
|
description: |
|
|
A temporal segment within a video, defined by start and end times.
|
|
|
|
**DEFINITION**:
|
|
|
|
VideoTimeSegment represents a bounded temporal portion of video content.
|
|
It is the foundational unit for time-coded content including:
|
|
- Subtitle/caption entries (text displayed at specific times)
|
|
- Annotation segments (detected scenes, objects, faces)
|
|
- Chapter markers (user-defined content sections)
|
|
|
|
**DUAL TIME REPRESENTATION**:
|
|
|
|
Times are stored in two formats for different use cases:
|
|
|
|
| Format | Example | Use Case |
|
|
|--------|---------|----------|
|
|
| ISO 8601 duration | PT0M30S | Human-readable, serialization |
|
|
| Seconds (float) | 30.0 | Computation, synchronization |
|
|
|
|
Both representations MUST be kept in sync. The seconds format is
|
|
primary for computation; ISO 8601 is derived for display/storage.
|
|
|
|
**MEDIA FRAGMENTS URI (W3C)**:
|
|
|
|
VideoTimeSegment aligns with W3C Media Fragments URI 1.0 specification
|
|
for addressing temporal fragments of video:
|
|
|
|
```
|
|
https://example.com/video.mp4#t=30,35
|
|
```
|
|
|
|
The `start_seconds` and `end_seconds` map directly to the `t=` parameter.
|
|
|
|
**WEB ANNOTATION COMPATIBILITY**:
|
|
|
|
When used as an annotation target selector:
|
|
- Maps to `oa:FragmentSelector` with `conformsTo` Media Fragments
|
|
- Enables interoperability with W3C Web Annotation Data Model
|
|
|
|
**CIDOC-CRM E52_Time-Span**:
|
|
|
|
In cultural heritage documentation:
|
|
- E52_Time-Span is the extent of a time-span
|
|
- Used for temporal properties of cultural objects
|
|
- VideoTimeSegment extends this to media-specific temporal segments
|
|
|
|
**CONFIDENCE SCORING**:
|
|
|
|
For segments generated by ASR (speech recognition) or CV (computer vision):
|
|
- `confidence`: 0.0-1.0 score for segment accuracy
|
|
- Enables filtering by quality threshold
|
|
- Critical for AI-generated transcripts and annotations
|
|
|
|
**HERITAGE USE CASES**:
|
|
|
|
| Use Case | Example | Start | End |
|
|
|----------|---------|-------|-----|
|
|
| Subtitle entry | "Welcome to the museum" | 0:30 | 0:35 |
|
|
| Scene annotation | "Exhibition hall panorama" | 1:00 | 1:30 |
|
|
| Chapter marker | "Introduction" | 0:00 | 2:00 |
|
|
| Object detection | "Painting: Night Watch" | 3:15 | 3:20 |
|
|
| Speaker change | "Curator speaking" | 5:00 | 7:30 |
|
|
|
|
exact_mappings:
|
|
- crm:E52_Time-Span
|
|
- oa:FragmentSelector
|
|
|
|
close_mappings:
|
|
- ma:MediaFragment
|
|
|
|
related_mappings:
|
|
- schema:Clip
|
|
|
|
slots:
|
|
# Time boundaries (ISO 8601 duration format)
|
|
- start_time
|
|
- end_time
|
|
|
|
# Time boundaries (seconds for computation)
|
|
- start_seconds
|
|
- end_seconds
|
|
|
|
# Content
|
|
- segment_text
|
|
- segment_index
|
|
|
|
# Quality
|
|
- confidence
|
|
|
|
# Metadata
|
|
- speaker_id
|
|
- speaker_label
|
|
|
|
slot_usage:
|
|
start_time:
|
|
slot_uri: ma:hasStartTime
|
|
description: |
|
|
Start time of segment as ISO 8601 duration from video beginning.
|
|
|
|
Media Ontology: hasStartTime for temporal start.
|
|
|
|
**Format**: ISO 8601 duration (e.g., "PT0M30S" = 30 seconds from start)
|
|
|
|
**Common Patterns**:
|
|
- PT0S = Start of video (0 seconds)
|
|
- PT30S = 30 seconds
|
|
- PT1M30S = 1 minute 30 seconds
|
|
- PT1H15M30S = 1 hour 15 minutes 30 seconds
|
|
range: string
|
|
required: false
|
|
pattern: "^PT(\\d+H)?(\\d+M)?(\\d+(\\.\\d+)?S)?$"
|
|
examples:
|
|
- value: "PT0M30S"
|
|
description: "30 seconds from video start"
|
|
- value: "PT1H15M30S"
|
|
description: "1 hour 15 minutes 30 seconds"
|
|
|
|
end_time:
|
|
slot_uri: ma:hasEndTime
|
|
description: |
|
|
End time of segment as ISO 8601 duration from video beginning.
|
|
|
|
Media Ontology: hasEndTime for temporal end.
|
|
|
|
Must be greater than or equal to start_time.
|
|
range: string
|
|
required: false
|
|
pattern: "^PT(\\d+H)?(\\d+M)?(\\d+(\\.\\d+)?S)?$"
|
|
examples:
|
|
- value: "PT0M35S"
|
|
description: "35 seconds from video start"
|
|
|
|
start_seconds:
|
|
slot_uri: hc:startSeconds
|
|
description: |
|
|
Start time in seconds (floating point) from video beginning.
|
|
|
|
**PRIMARY for computation**. Use for:
|
|
- Video player synchronization
|
|
- Duration calculations
|
|
- Time-based sorting and filtering
|
|
|
|
Precision to milliseconds (3 decimal places) is typical.
|
|
range: float
|
|
required: true
|
|
minimum_value: 0.0
|
|
examples:
|
|
- value: 30.0
|
|
description: "30 seconds from start"
|
|
- value: 30.500
|
|
description: "30.5 seconds (millisecond precision)"
|
|
|
|
end_seconds:
|
|
slot_uri: hc:endSeconds
|
|
description: |
|
|
End time in seconds (floating point) from video beginning.
|
|
|
|
Must be greater than start_seconds.
|
|
|
|
For single-frame annotations (e.g., object detection in one frame),
|
|
end_seconds may equal start_seconds or be slightly greater.
|
|
range: float
|
|
required: true
|
|
minimum_value: 0.0
|
|
examples:
|
|
- value: 35.0
|
|
description: "35 seconds from start"
|
|
|
|
segment_text:
|
|
slot_uri: oa:bodyValue
|
|
description: |
|
|
Text content for this segment.
|
|
|
|
Web Annotation: bodyValue for textual content.
|
|
|
|
**Usage by content type**:
|
|
- Subtitles: Displayed caption text
|
|
- Transcripts: Spoken words during this segment
|
|
- Annotations: Description of detected content
|
|
- Chapters: Chapter title/description
|
|
range: string
|
|
required: false
|
|
examples:
|
|
- value: "Welkom bij het Rijksmuseum"
|
|
description: "Dutch subtitle text"
|
|
- value: "The curator explains the painting's history"
|
|
description: "Transcript segment"
|
|
|
|
segment_index:
|
|
slot_uri: hc:segmentIndex
|
|
description: |
|
|
Sequential index of this segment within the parent content.
|
|
|
|
Zero-based index for ordering segments:
|
|
- Subtitle: Order in which captions appear
|
|
- Annotation: Detection sequence
|
|
|
|
Enables reconstruction of segment order when times overlap
|
|
or for stable sorting.
|
|
range: integer
|
|
required: false
|
|
minimum_value: 0
|
|
examples:
|
|
- value: 0
|
|
description: "First segment"
|
|
- value: 42
|
|
description: "43rd segment (zero-indexed)"
|
|
|
|
confidence:
|
|
slot_uri: hc:confidence
|
|
description: |
|
|
Confidence score for AI-generated content.
|
|
|
|
Range: 0.0 (no confidence) to 1.0 (complete certainty)
|
|
|
|
**Applies to**:
|
|
- ASR-generated transcript/subtitle segments
|
|
- CV-detected scene or object annotations
|
|
- OCR-extracted text from video frames
|
|
|
|
**Thresholds** (suggested):
|
|
- > 0.9: High confidence, suitable for display
|
|
- 0.7-0.9: Medium, may need review
|
|
- < 0.7: Low, flag for human verification
|
|
range: float
|
|
required: false
|
|
minimum_value: 0.0
|
|
maximum_value: 1.0
|
|
examples:
|
|
- value: 0.95
|
|
description: "High confidence ASR segment"
|
|
- value: 0.72
|
|
description: "Medium confidence, may contain errors"
|
|
|
|
speaker_id:
|
|
slot_uri: hc:speakerId
|
|
description: |
|
|
Identifier for the speaker during this segment.
|
|
|
|
For transcripts with speaker diarization:
|
|
- Links to identified speaker (e.g., "SPEAKER_01")
|
|
- May be resolved to actual person identity
|
|
|
|
Enables multi-speaker transcript navigation.
|
|
range: string
|
|
required: false
|
|
examples:
|
|
- value: "SPEAKER_01"
|
|
description: "First identified speaker"
|
|
- value: "curator_taco_dibbits"
|
|
description: "Resolved speaker identity"
|
|
|
|
speaker_label:
|
|
slot_uri: hc:speakerLabel
|
|
description: |
|
|
Human-readable label for the speaker.
|
|
|
|
Display name for the speaker during this segment:
|
|
- May be generic ("Narrator", "Interviewer")
|
|
- May be specific ("Dr. Taco Dibbits, Museum Director")
|
|
|
|
Distinguished from speaker_id which is a machine identifier.
|
|
range: string
|
|
required: false
|
|
examples:
|
|
- value: "Narrator"
|
|
description: "Generic speaker label"
|
|
- value: "Dr. Taco Dibbits, Museum Director"
|
|
description: "Specific identified speaker"
|
|
|
|
rules:
|
|
- postconditions:
|
|
description: end_seconds must be >= start_seconds
|
|
# Note: LinkML doesn't support direct comparison rules,
|
|
# but this documents the constraint for validation
|
|
|
|
comments:
|
|
- "Reusable time segment for subtitles, annotations, chapters"
|
|
- "Dual time format: ISO 8601 for serialization, seconds for computation"
|
|
- "Aligns with W3C Media Fragments URI specification"
|
|
- "Confidence scoring for AI-generated content"
|
|
- "Speaker diarization support for multi-speaker transcripts"
|
|
|
|
see_also:
|
|
- "https://www.w3.org/TR/media-frags/"
|
|
- "https://www.w3.org/TR/annotation-model/"
|
|
- "https://www.w3.org/ns/ma-ont"
|
|
- "http://www.cidoc-crm.org/cidoc-crm/E52_Time-Span"
|
|
|
|
# ============================================================================
|
|
# Slot Definitions
|
|
# ============================================================================
|
|
|
|
slots:
|
|
start_time:
|
|
description: Start time as ISO 8601 duration from video beginning
|
|
range: string
|
|
|
|
end_time:
|
|
description: End time as ISO 8601 duration from video beginning
|
|
range: string
|
|
|
|
start_seconds:
|
|
description: Start time in seconds (float) from video beginning
|
|
range: float
|
|
|
|
end_seconds:
|
|
description: End time in seconds (float) from video beginning
|
|
range: float
|
|
|
|
segment_text:
|
|
description: Text content for this time segment
|
|
range: string
|
|
|
|
segment_index:
|
|
description: Sequential index of segment within parent
|
|
range: integer
|
|
|
|
confidence:
|
|
description: Confidence score for AI-generated content (0.0-1.0)
|
|
range: float
|
|
|
|
speaker_id:
|
|
description: Identifier for speaker during this segment
|
|
range: string
|
|
|
|
speaker_label:
|
|
description: Human-readable label for speaker
|
|
range: string
|