Multi-Modal Multi-Task Foundational Models for Medical Image Manipulation

ChangeBridge: Patent Apps - Health Informatics (G16H)

Published April 2nd, 2026

Detected April 2nd, 2026

Summary

The USPTO published patent application US20260094693A1 on April 2, 2026, disclosing a multi-modal AI system developed by inventors Alexandru Constantin Serban, Mehmet Akif Gulsun, Puneet Sharma, and Dorin Comaniciu for medical image manipulation and information retrieval. The system receives text-based instructions, encodes them using machine learning text encoders, and determines and performs instructions through medical applications to generate responses. The publication affects medical device makers, healthcare technology providers, and AI developers working on clinical applications.

View original document View source feed page

What changed

The USPTO published patent application US20260094693A1 disclosing systems and methods for multi-modal multi-task foundational models that process text-based instructions to perform actions on medical applications. The technology uses machine learning text encoder networks to convert text instructions into features, a policy module to determine executable instructions, and medical applications to perform those instructions and generate responses for medical image manipulation and information retrieval.

No immediate compliance action is required as this is a patent publication rather than a regulatory requirement. However, medical device manufacturers, healthcare technology companies, and AI developers should review the patent claims to assess potential IP landscape implications for their own medical AI products. Companies developing similar text-instructed medical imaging systems may need to evaluate freedom-to-operate considerations.

Source document (simplified)

← USPTO Patent Applications

MULTI-MODAL MULTI-TASK FOUNDATIONAL MODELS FOR MEDICAL IMAGE MANIPULATION AND INFORMATION RETRIEVAL

Application US20260094693A1 Kind: A1 Apr 02, 2026

Inventors

Alexandru Constantin Serban, Mehmet Akif Gulsun, Puneet Sharma, Dorin Comaniciu

Abstract

Systems and methods for automatically performing one or more actions on one or more medical applications are provided. Text-based instructions are received. The text-based instructions are encoded into text features using a machine learning based text encoder network. One or more instructions for performing by one or more medical applications are determined using a policy module based on the text features. The one or more instructions are performed by the one or more medical applications to generate a response to the text-based instructions. The response to the text-based instructions is output.