Care / Medical / Monitoring Image Analysis

AI-sCare+ Care Monitoring Image Analytics System

Night patrol and room checks were burdensome, while the facility needed earlier signs of abnormal events, privacy-aware operation and record management.

AI画像解析介護見守りWebRTCEdge AIイベント録画

Results Summary

1

Stable inspection quality: AI-based judgment dramatically reduces human error

2

Reduced workload: Automated recording and assessment cuts labor hours

3

Fast PoC-to-production: Evaluation kit enables rapid validation and smooth transition

Customer Challenge

Night patrol and room checks were burdensome, while the facility needed earlier signs of abnormal events, privacy-aware operation and record management.

Implementation Overview

Background

Care facilities must balance resident safety and staff workload. Night patrol, records and emergency judgment can place heavy pressure on operations.

Challenges

  • Room-by-room checks took time
  • It was difficult to review context before and after events
  • Video operation had to respect privacy rules
  • The system needed phased expansion by facility scale

Implemented Solution

Room cameras, edge AI inference, WebRTC viewing, event recording and management UI were combined, with notification rules tuned to workflow.

  • Low-light cameras
  • Edge AI inference device
  • Event notification and recording
  • Management terminal and access control

Rollout

  1. Pilot Area
    Room environment and workflow were reviewed to define a small PoC.
  2. Detection Tuning
    AI and thresholds were tuned to lighting, bed layout and notification criteria.
  3. Operation Review
    Notification, video review and record workflow were checked with staff.

Operational Change

  • Patrol check points became clearer and helped staff notice state changes
  • Event context became easier to review, supporting standardized records
  • A structure suitable for phased expansion was established

Expansion

Room expansion, additional notification targets and links with nurse-call or record systems can be reviewed step by step.

Discuss a Similar Facility

Share facility scale and current issues if you would like a PoC designed around your room environment and workflow.

Results & Impact

Event detection, remote viewing and record review were organized so staff could judge response priority more easily.

Why CSUN

1

Integrated camera, embedded, AI, and connectivity: System-level optimization, not piecemeal

2

One-stop from PoC to mass production: Prototype evaluation → design → production → long-term supply

3

Flexible design for real-world constraints: Adapts to existing lines, lighting conditions, and space limits

4

Trilingual support (JP/ZH/EN): Global development, sourcing, and support capabilities

Technical Details

Deployment FieldCare, Healthcare & Monitoring
Monitoring TargetsElderly residents (fall, wandering, sudden health changes), facility residents
SensorsAI Cameras (RK3399+OV4689), biometric sensors, temperature/humidity, toilet occupancy
VideoWebRTC P2P live video, local recording (NAS/SD card)
AI ProcessingEdge AI (RK3399), CNN pose estimation, fall detection, wandering detection
NotificationInstant alerts to iPad/PC (optional cloud notifications)
Client DevicesiPad, iPhone, Android, Windows PC
RecordingLocal recording, video search & download, NAS/SD card support
PrivacyEdge processing + blurring, cloud-independent architecture
ResultNight round reduction, false-positive mitigation, automated care record integration, faster incident review
Deployment Track RecordDeployed across 8 facilities and 700 rooms

Related Solutions & Products

From camera selection to AI model design, embedded implementation, and mass production — we provide end-to-end support.

Discuss a Similar Case

FAQ

Can this work in a similar production environment?

In most cases, yes. Please share sample images and requirements from your site. We'll propose the optimal configuration based on similar cases.

Can I consult just for a PoC?

Absolutely. We welcome PoC-only engagements. Our evaluation kit enables rapid validation.

Can we validate without stopping our existing line?

Yes. Add-on cameras or non-invasive sensors allow validation while your existing line keeps running.

Can you help with camera selection?

Yes. We'll recommend the optimal camera module based on your target objects, lighting conditions, and inspection criteria.

Facing a similar challenge?

From camera selection to AI model design, embedded implementation, and mass production — we provide end-to-end support.