Intellectual Property India Publishes Patent Application for A Machine Learning-Based Intelligent System For Automated Autism Spectrum Disorder Detection

MUMBAI, India, Oct. 12 -- Intellectual Property India has published a patent application (202641119650 A) filed by Dr K. Anup Kumar; Ms. Swapna Rani G; Dr Dhanavath Nagaraju; Ms. Marthineni Shilpa; Dr. P. Jyothi; Dr. P. Ashwini; Mrs. G. Rani; Mrs. J. Rachana; Mrs. Sangeeta Jawar; and Mr. Rama Krishna Guduru on October 06, 2026, for A Machine Learning-Based Intelligent System For Automated Autism Spectrum Disorder Detection.

Inventors include Dr K. Anup Kumar; Ms. Swapna Rani G; Dr Dhanavath Nagaraju; Ms. Marthineni Shilpa; Dr. P. Jyothi; Dr. P. Ashwini; Mrs. G. Rani; Mrs. J. Rachana; Mrs. Sangeeta Jawar; and Mr. Rama Krishna Guduru.

The application for the patent was published on October 09, 2026, under issue no. 41/2026.

Abstract: The present system relates to a Machine Learning-Based Intelligent System for Automated Autism Spectrum Disorder (ASD) Detection. The system provides an intelligent multimodal framework for automated screening of ASD-associated behavioral and communication patterns using behavioral questionnaire data, speech and language characteristics, facial expressions, eye-gaze information, and video-based behavioral features. The proposed system employs data preprocessing, feature engineering, Principal Component Analysis (PCA) for dimensionality reduction, and an Adaptive Multimodal Feature Fusion (AMFF) technique to integrate heterogeneous assessment data. The fused feature representation is analyzed using a hybrid machine-learning classification technique comprising Support Vector Machine (SVM), Random Forest (RF), and Gradient Boosting/XGBoost, with an optional deep-learning model for extracting complex visual and temporal features. An Explainable Artificial Intelligence (XAI) technique such as SHAP is employed to identify the contribution of individual features and data modalities, while a confidence estimation mechanism evaluates prediction reliability. The expected output comprises an ASD-associated risk score, classification result such as Low-Risk, Moderate- Risk, or High-Risk, prediction confidence score, modality-wise contributing features, and an automated assessment report indicating whether further professional clinical evaluation is recommended. The system further supports missing-data handling and longitudinal monitoring, thereby providing an interpretable and scalable computational tool for ASD-associated screening and professional decision support.

Disclaimer: Curated by HT Syndication.