AutiSmart
AI-based ASD detection and therapy support platform for early and middle-stage intervention workflows.
Public Safety Note
This page intentionally presents a safe, high-level case study. Sensitive datasets, internal implementation details, and private source code are intentionally excluded.
Project Overview
AutiSmart is a full-stack Final Year Project (FYP) developed to support early and middle-stage Autism Spectrum Disorder (ASD) detection and therapy. The platform combines AI-driven multimodal analysis with practical care workflows so experts and caregivers can assess progress, coordinate interventions, and track outcomes in one system.
Problem Statement
ASD assessment and therapy support are often fragmented across manual records, disconnected channels, and non-standard reporting. This makes early-stage decision-making difficult and slows therapy adaptation. AutiSmart consolidates assessment, therapy planning, communication, and progress reporting into a unified role-based platform.
Core Features
- Multimodal AI Support: Audio, video, and text analysis pipelines used to assist ASD stage suggestion workflows
- Emotion Recognition: Computer vision-driven emotion signals integrated into monitoring and therapy context
- Adaptive Therapy Plans: Dynamic content generation through APIs to personalize recommendations by profile and progress
- Interactive Therapy Games: Engagement-focused game modules to support therapy participation and continuity
- Expert-Caregiver Communication: Real-time coordination channels between professionals and caregivers
- Automated Progress Tracking: Continuous activity capture with report generation for review and follow-up
- Role-Based Security: Authenticated multi-user workflows with controlled access to sensitive records
Technology Stack
Frontend
React.js with role-aware interfaces, responsive design, and dashboard-style therapy monitoring views
Backend
Node.js with secure API architecture, authentication, and modular services for assessment and reporting workflows
Database
MongoDB with structured collections for user roles, therapy plans, progress records, and generated summaries
AI and CV Integration
Multimodal analysis support (audio/video/text), computer vision based emotion recognition, and dynamic recommendation APIs
Architecture Decisions
- Role-aware access model keeps caregiver, expert, and administration workflows separated and traceable
- Assessment support, therapy planning, communication, and reporting are modularized for maintainability
- Progress capture and report generation are designed for longitudinal review and practical follow-up
- Privacy-safe publication model exposes workflow and outcomes while protecting sensitive implementation detail
Real-World Impact
- Demonstrated a full-stack approach for ASD support from assessment context to therapy follow-up
- Improved coordination readiness between experts and caregivers through shared workflow visibility
- Showcased practical AI-assisted feature integration in a healthcare-oriented university project
- Delivered an auditable progress-and-reporting model suitable for iterative therapy planning
Repository Status
Private repository maintained for project safety and responsible disclosure. Public documentation is intentionally limited to non-sensitive outcomes and workflow context.