{"id":56104,"date":"2025-12-15T08:51:09","date_gmt":"2025-12-15T08:51:09","guid":{"rendered":"https:\/\/www.bravent.net\/?p=56104"},"modified":"2025-12-15T08:51:10","modified_gmt":"2025-12-15T08:51:10","slug":"meta-launches-sam-3-the-new-model-capable-of-finding-and-tracking-almost-anything-in-images-and-video","status":"publish","type":"post","link":"https:\/\/www.bravent.net\/en\/news\/meta-launches-sam-3-the-new-model-capable-of-finding-and-tracking-almost-anything-in-images-and-video","title":{"rendered":"Meta launches SAM 3: the new model capable of finding and tracking \u201calmost anything\u201d in images and video"},"content":{"rendered":"

Meta<\/a> has announced the open release of SAM 3<\/a> (Segment Anything Model 3)<\/strong>, the most advanced evolution of its segmentation model family.<\/p>\n

This model redefines how objects are identified and tracked in images and video: now a simple text prompt or example image is enough for the system to locate all instances that match the description.<\/p>\n<\/div>[vc_empty_space]<\/div><\/div><\/div><\/section>

Unlike previous versions, which required clicks or bounding boxes to delimit objects, SAM 3 understands natural language, enabling instructions such as \u201cpeople with backpacks,\u201d \u201cred cars,\u201d or \u201cbicycles on the right.\u201d<\/p>\n

The model also maintains stable temporal consistency, allowing it to follow moving objects even when the shot or lighting changes.<\/p>\n<\/div>[vc_empty_space][vc_single_image image=”56117″ img_size=”Full” alignment=”center”][vc_empty_space]<\/div><\/div><\/div><\/section>

What makes SAM 3 different<\/strong><\/h2>\n

\ud83d\udd39 Text- or image-guided segmentation<\/strong>
\nYou can describe what you want to find or upload a reference image. SAM 3 automatically identifies all matches.<\/p>\n

\ud83d\udd39 Video tracking<\/strong>
\nIt generates consistent masks across frames\u2014essential in dynamic environments.<\/p>\n

\ud83d\udd39 Detection + tracking in a single system<\/strong>
\nOne model performs both tasks, simplifying computer-vision pipelines.<\/p>\n

\ud83d\udd39 Fully open source<\/strong>
\nMeta released the code and weights, enabling research, deployment, and enterprise integrations.<\/p>\n<\/div>[vc_empty_space][vc_single_image image=”56111″ img_size=”Full” alignment=”center”][vc_empty_space]<\/div><\/div><\/div><\/section>

Real-world use cases and applications of SAM 3<\/strong><\/h2>\n
    \n
  1. Counting people in queues or events<\/strong>
    \nIdentifies and counts individuals without custom models.<\/li>\n
  2. Tracking moving vehicles<\/strong>
    \nHighly useful for mobility, access control, traffic analytics, or smart parking.<\/li>\n
  3. Automatic face blurring<\/strong>
    \nEnsures privacy in corporate videos, public recordings, or datasets.<\/li>\n
  4. Automatic license-plate blurring<\/strong>
    \nRegulatory compliance with no manual work.<\/li>\n
  5. Smart editing and post-production<\/strong>
    \nManipulate objects with prompts like \u201cblue chair\u201d without manual cutouts.<\/li>\n
  6. Retail and e-commerce<\/strong>
    \nProduct detection, shelf analytics, and interactive experiences.<\/li>\n
  7. Science, sports, and technical analysis<\/strong>
    \nPrecise tracking of animals, particles, athletes, or machinery.<\/li>\n<\/ol>\n<\/div>[vc_empty_space][vc_single_image image=”56113″ img_size=”Full” alignment=”center”][vc_empty_space]<\/div><\/div><\/div><\/section>

    Availability: SAM 3 is now on Azure Microsoft Foundry<\/strong><\/h2>\n

    Microsoft has added facebook-sam3<\/em> to the official Microsoft Foundry<\/a> model catalog, accessible from the Hugging Face registry.<\/p>\n<\/div>[vc_empty_space][vc_single_image image=”56113″ img_size=”Full” alignment=”center”][vc_empty_space]

    This means companies can use SAM 3 directly on Azure, with no need to deploy it manually in containers or on their own GPUs.<\/p>\n<\/div>[vc_empty_space]<\/div><\/div><\/div><\/section>

    Official resources<\/strong><\/h2>\n