Social media has a high effect on our everyday lives. Peoples share their views, stories, news, and communicate events through internet based life. It leads to the huge shared information in the social media. It is not convenient to find and frame the essential events with the huge data, in most of the cases, browsing, searching, and monitoring events turns out to be very challenging. Major work has been done on topic detection and tracking (TDT) domain. Many of such methods are on the basis of single-modality (e.g., text, images) or multi-modality data. In the analysis of single-modality, various available methods acquire visual data (e.g., images and videos) or text based data (e.g., names, time references, locations, title, tags, and description) separately to design the event data for the event detection and tracking as well. This issue can be cleared by new multi-modal social event tracking and a transformative system for effectively capturing the events, as well as make the event synopsis in time. The authors present a novel method that works with the mmETM, which can viably make the social records, and it includes the extensive content incorporated with pictures. To coordinate the approach for social tracking, an incremental learning approach is obtained like mmETM that gives information and the event's visual topics in social media. To support this work, the authors have utilized an example informational index and regulated a few tests on it. Both the qualitative and the quantitative investigation on the proposed mmETM approach have shown some best in-class strategies.