The ISOCELL Zoom Anywhere for 200MP Camera from Samsung Is Stunning


A few modifications to Samsung’s 200-megapixel image sensor—which is rumored to be found in the future Galaxy S24 Ultra—were revealed this morning. This sensor is presently found in the Galaxy S23 Ultra, but Samsung unveiled new solutions for it this week. These solutions are driven by Qualcomm’s new Snapdragon 8 Gen 3 processor, which focuses on artificial intelligence.

One of the new options is ISOCELL Zoom Anywhere, which lets device owners choose a topic and the camera will track and record it in high definition automatically. With this function, ideally, there won’t be as many wobbly movies as there usually are when filming at close range. Samsung states that the sensor’s improved real-time remosaic capabilities and Qualcomm’s AI engine enable this feature.

Speaking of remosaic capabilities, I’ll let Samsung explain as it takes a NASA scientist to comprehend what is going on.

Currently, the steps the 200MP image sensor follows for processing images after shooting are: raw data output occurs before a remosaic, which then leads to image signal processing (ISP) before finishing with JPEG image output. In this process, all of the steps occur in sequence.

E2E AI Remosaic improves the process by enabling the remosaic and ISP to take place simultaneously. In other words, E2E AI Remosaic takes a process that occurs in sequence and makes it occur in parallel, thereby reducing remosaic latency by up to half. The end result is an overall faster image processing time, which reduces the shooting time of 200MP images and improves image quality.

Owners of the S24 Ultra can also anticipate the capability of concurrently recording videos in full field of view and at a high magnification. Because both videos were shot in high definition, there was greater flexibility during the shooting process. Once more using the scientific explanation, Samsung highlights the significance of this feature in relation to their remosaicing research.

Thanks to in-sensor zoom by Tetra pixel, users don’t need to be selective because of zoom limitations, either. Previously, with digital zoom, if the center region of interest (ROI) area was zoomed in 4x at a 12.5MP sum mode, it theoretically resulted in 0.78MP (12.5MP/16) image resolution. With in-sensor zoom, however, the selected area to be zoomed in on is selected from the entire scope and remosaiced, which allows the sensor to maintain the original 12.5MP resolution and enable video to be shot much more vividly.

And get notified everytime we publish a new blog post.