Features & Roadmap | Singapore Government Developer Portal
features roadmap
Overview
Features & Roadmap
Getting Started
Resources
Meet The Team
dev-portal-icon / PRODUCTS / analytics / vas / features roadmap

Features & Roadmap

Features

Media Hub (Media Management and Sharing Services)

  • Video Upload, Playback, Search, Request and Sharing
  • Integration with Cloud Video Exchange (CVX) for video import from onboarded camera sources for analytics
  • Image dataset upload, management and sharing

VideoIO (VA Enablement Service)

  • Annotate images and videos, and develop object detection models through an intuitive UI
  • Train video analytics models using cloud resources
  • Manage multiple projects and multiple classes
  • Use less data and prototype faster with transfer learning and data augmentation
  • Compatible with multiple image formats

Insights Service (VA-as-a-Service)

  • Generate insights for your videos based on ready-to-use VA services for detection and counting of relevant events and objects
  • VA Services that are currently available include:
    • Smoking action detection and count (Balefire)
    • Detection of Personal Mobility Devices (PMDs), pedestrians and cyclists (Mobius)
    • Number plate recognition and detection of parking behavior by unique vehicles, with identification of vehicular types (PlateSnap)
    • Detection and dwell time estimation of marine vessels and kayaks (Seafari)
    • Biometric photo compliance check (LISA)
    • Detection of entities and groups of entities; estimation of distance between entities to provide notification of clustered groups where entities are closely spaced (TorusVision)
    • Prompt-based general object detection
  • Allows customisation of regions of interest (for supported services)
  • Users can generate shareable reports, with automatic notifications when the analysis has been completed
  • API support for selected image analytics services
  • Export reports in CSV format for custom analysis

Video Indexer

  • Extracts rich metadata from videos, such as:
    • Topics and conceptual entities
    • Object and action identification and labelling
    • Different scenes with reference keyframe extraction
    • Visual text recognition
    • Transcription
  • Search across videos and within videos for specific topics or tags
  • Navigate easily to segments within the video containing the topic of interest

Video Summary

  • Provide a summarized view of video events for fast review
    • Detect a rich variety of object
    • Allow advanced filtering options on detected objects for attributes such as type, colour and more
    • Search within videos for based on image similiarity
  • Determine movement and direction trends with rich overlays
  • Allows users to automatically zero in on events within videos for analysis

Roadmap

The team will continue to enhance and launch new capabilities with the aim to provide seamless content management experience and help agencies get the most out of their content:

  • Enhanced media management experience for WOG agencies by bringing together smart content analysis, enabling better search and discovery, and content metadata tagging
  • Ability to test and customize approaches to derive insights from media content utilizing foundational vision models
  • Integration with Cloud Video Exchange (CVX) enabling WOG users to import videos and images from their camera systems and derive actionable insights anywhere, anytime.
  • Insights automations that will allow users to setup recurring analytics workflows with easy-to-use interfaces and notification alerts

Techstack

TypeScript + React, NestJs, NodeJS, Python, Major AWS Services (EC2, Sagemaker, Fargate, EC2ECS, RDS, Lambda), Terraform, Docker, CDK

Was this article useful?

A Centralised Platform That Streamlines Video Analytics Integration Across Government Agencies