MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641100698 A) filed by Jeppiaar Institute Of Technology on August 20, 2026, for A Multimodal Artificial Intelligence Framework For Real-Time Detection And Verification Of Synthetic.
Inventors include Mr. Saravanan S; Shalini B; Avinashraj K S; Hemaraj; Dinesh Karthik; and Vishal B.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: 6.ABSTRACT OF THE INVENTION (to be given along with complete specification on separate page) The rapid growth of Generative Artificial Intelligence has made it possible to create highly realistic fake images, videos, and audio. Although these technologies have many useful applications, they can also be misused to spread misinformation, impersonate individuals, commit fraud, and manipulate digital content. Detecting such synthetic media is becoming increasingly difficult because many existing detection systems mainly focus on a single type of media, such as images or video frames. This can limit their ability to identify advanced manipulations that involve multiple forms of media. This project proposes a Multimodal Artificial Intelligence Framework for Real-Time Detection and Verification of Synthetic Image, Video, and Audio Content Using Dynamic Cross-Modal Evidence Fusion. The proposed system examines images, videos, and audio using different Artificial Intelligence and Deep Learning techniques and combines the results to determine whether the content is genuine or manipulated. The image analysis component identifies visual and facial abnormalities, while the video component examines spatial and temporal inconsistencies. The audio component analyzes speech patterns and other characteristics that may indicate artificially generated or modified audio. In addition to analyzing each modality separately, the system checks whether the audio, visual, and temporal information are consistent with one another. For example, it can identify situations where a person's lip movements do not match the speech or where visual and audio information show signs of manipulation. The system uses a dynamic evidence fusion mechanism, which gives different importance to each modality depending on its quality, reliability, and detection confidence. The framework produces an overall authenticity score along with individual confidence scores for the image, video, and audio components. It can also provide understandable evidence about the possible reasons for identifying content as synthetic or manipulated. Deep learning models such as Convolutional Neural Networks, Vision Transformers, temporal models, and audio-processing models can be integrated into the framework. Overall, the proposed system aims to provide a reliable, scalable, and explainable solution for real-time verification of digital media across multiple modalities
Disclaimer: Curated by HT Syndication.
