AI-Enhanced Telemedicine System for Divyangjan: A Context-Sensitive Framework for Medication Management and Healthcare Accessibility

Abstract

Ensuring equitable healthcare access for individuals with visual disabilities (Divyangjan) presents persistent obstacles, especially regarding pharmaceutical management and remote medical consultation. This work introduces an integrated framework combining artificial intelligence with telemedicine infrastructure specifically engineered for users experiencing vision loss. The framework incorporates vision-based pharmaceutical recognition utilizing advanced Inception-v4 architectures, demonstrating 97.5% classification performance under standardized conditions and 94.2% accuracy across variable real-world scenarios. Language understanding capabilities leverage transformerbased architectures fine-tuned on medical conversation datasets, achieving 92% intent classification. Accessibility features span multiple modalities: auditory guidance with 95% speech command accuracy, tactile response mechanisms, and optimized audio interfaces. An adaptive personalization system employing reinforcement learning techniques modifies interface behavior according to individual user characteristics, ambient conditions, and usage patterns, yielding 30-40% gains in user contentment metrics. Hospital-based evaluation involving 45 participants with visual impairments validated 96% pharmaceutical identification success across heterogeneous medication samples, achieving mean completion intervals of 78 seconds versus 180 seconds for conventional approaches. The platform attains 92% mean task success and 4.6/5.0 satisfaction scores. Connectivity with health information systems, pharmaceutical repositories, and urgent care infrastructure supports holistic care coordination. Findings suggest potential for 35-45% enhancement in medication compliance and 25-35% decrease in emergency intervention latency.

Ähnliche Arbeiten