Using machine learning and image recognition for automatic relocation of camera display area and sizing of camera image
AssigneeSony Interactive Entertainment Inc.
InventorNoam Rimon
Life of this patent
Estimating the term…
Abstract
Machine learning (ML) and image recognition are used to identify an area in a presentation such as a slide presentation which is not occupied by important content. Discovery is made of background images or data that is common between multiple slides, which implies that such common content is a banner or other element that is not unique to the current slide, in turn implying that it is not critical content. In this way, images from a camera sought to be overlaid on the presentation can be automatically relocated and, if desired, resized on a slide by slide basis to areas of the non-critical content. Manual adjustment can be used for reinforcement learning of ML for future slides.
Claims18
A device comprising:
- at least one processor configured with instructions to:
- identify that content in a first area in a first slide of a computerized presentation is substantially identical to content in a second area of a second slide of the computerized presentation, the first and second areas respectively occupying substantially a same geometric region of the first and second slides, the content being visual alpha-numeric content, graphical content, or both alpha-numeric and graphical content;
- modify an image to fit within the second area of the second slide to render a modified image;
- overlay the modified image into the second area of the second slide;
- identify a region of a projected slide as containing pointer light;
- and
- responsive to identifying the region of the projected slide as containing pointer light, not overlay the modified image into an area of the slide corresponding to the region.
A computerized method, comprising:
- identifying at least a first area of a first computerized presentation element as being a candidate for having an image overlaid thereon, the first area being a subset of an area of the first presentation element;
- and
- responsive to identifying the first area of the first computerized presentation element as being a candidate for having an image overlaid thereon, overlaying an image on the first area, wherein the method further comprises:
- identifying a region of a projected slide as containing pointer light;
- and
- responsive to identifying the region of the projected slide as containing pointer light, not overlaying the image into an area of the first presentation element corresponding to the region.
An apparatus, comprising:
- at least one processor programmed with instructions to:
- discover content that is common between multiple slides, the content being alpha-numeric, graphical, or both alpha-numeric and graphical;
- based at least in part on discovering content that is common between multiple slides, automatically relocate, or resize, or relocate and resize at least one image from a camera on a slide by slide basis to areas of non-critical content;
- identify a region of a projected slide as containing pointer light;
- and
- responsive to identifying the region of the projected slide as containing pointer light, not overlaying an image into an area of a slide corresponding to the region.
Dependencies are inferred from the claim text.