Microsoft Technology Licensing, LLC

Tomasz S M Kasperkiewicz

Steven J White

David L Albert

Richard S Szeliski

Blaise Aguera Y Arcas

Thomas W Olsen

David R Parlin

Richard S Turner, Jr

Robert Albert Wlodarczyk

Eric W Richards

Eric S Vandenberg

Timothy R O'Connor

Donald M Marsh

Richard Shailer Turner, Jr

Benjamin R Peart

Roger A Bischoff

David R Blythe

Scott E Dart

Jordan L K Schwartz

Jingyang Xu

Po Yuan

Alexander Sabo Brodie

Peter Alexi Gurevich

Robert E Sinclair, Ii

Mahmood Gulam Qadir

Steve J White

Kathleen M Frigon

Eric J Richards

Timothy R O'CONNOR

Images are analyzed within a 3D environment that is generated based on spatial relationships of the images and that allows users to experience the images in the 3D environment. Image analysis may include ranking images based on user viewing information, such as the number of users who have viewed an image and how long an image was viewed. Image analysis may further include analyzing the spatial density of images within a 3D environment to determine points of user interest.

Analysis of images located within three-dimensional environments

A determination is made for each of multiple regions in multiple images of how good that region is perceived as being. A base image is identified, and a combined image is generated from the multiple images by automatically replacing each region of the base image with a corresponding region of another image if the corresponding region has been determined as being better than the region of the base image. The generating of the combined image can include automatically selecting from one of the multiple images a region in which an object that is present in one or more corresponding regions of other images is absent. Additionally, for a particular region of the base image, corresponding regions of the other images can be displayed, and the particular region replaced with a user-selected one of the corresponding regions of the other images.

Generating a combined image from multiple images

A system, a method and computer-readable media for processing raw image data with a graphics processing unit (GPU). Raw image data generated by an imaging sensor is received. A set of instructions for demosaicing the raw image data is communicated to the GPU. The GPU is enabled to demosaic the raw image data by executing the set of instructions.

Raw image processing

Topic-centric-media sharing allows a user to search for and share media objects related to a topic. The user can select media objects stored on a client computing device to share with other users. The client computing device contains software that, in operation, searches for a set of media objects related to the same topic as the selected media objects. If such a set is located, the software on the client computing device can associate the selected media objects with the set. If no such set is located, the user can create a new set to share with other users.

Topic centric media sharing

An arrangement is provided for automatically correcting a digital image by first examining the image to determine the class or type of correction that is needed to produce a more accurate rendition. The classes or types of corrections may include such items as exposure, color, depth of field, sharpness, distortion, and combinations thereof. The type of correction that is needed by any particular digital image is determined from an analysis of the image itself. Instead of presenting the user with a single corrected image, the user is typically presented with different renditions of the image. Each rendition corrects for the particular type of correction that has been identified, but in different ways from one another.

Automatic image correction providing multiple user-selectable options

A system and method for dynamically modifying a media presentation is disclosed. A first presentation theme associated with a media presentation is determined. A plurality of multi-media data files are accessed, wherein each of the plurality of multi-media data files comprises associated metadata. A first of the plurality of multi-media data files are associated with the first presentation theme based on the associated metadata and the first of the plurality of multi-media data files is presented.

Dynamically modifying a theme-based media presentation

Network-based and client-based operating environments implement computerized systems and techniques for organizing and discovering digital media objects originating from personal digital media sources. The association between a user-selected computer-usable descriptor and a particular digital media object serves as a semantic medium for exchange of information about the digital media object via a network, independent of whether the digital media object is registered with an established digital content sharing service. A user-selected computer-usable descriptor is part of one or more collections of such descriptors that are organized into published data structures having data elements corresponding to particular real-world meanings. Leveraging the consistent associations of computer-usable descriptors with the subject matter of personal digital media content within digital media objects enables the widespread and efficient computerized discovery of relevant digital media objects.

Taxonometric personal digital media organization

Applications that extend the basic concept of determining spatial relationships between images and generating a model allowing users to experience the images in a 3D environment are provided. One application is directed to the creation of documents that may be readily shared, viewed, and edited, wherein the documents include references to underlying content comprising a 3D environment and images positioned therein and captures connections between images. Another application relates to ranking images based on user viewing information, such as the number of users who have viewed an image and how long an image was viewed. Yet another application is directed to analyzing the spatial density of images within a 3D environment to determine points of user interest. A further application relates to using spatially indexed images to navigate between webpages and 3D environments.

Applications of three-dimensional environments constructed from images

A method and architecture for obtaining a desired frame and identifying a selected frame in an image container. A request for a frame of an image is received. The request may include an identifier that identifies a type of frame. In response thereto, a frame is located corresponding to the identifier and a reference to the frame is provided. In another aspect, a request is received to identify a frame of an image. A determination is made as to what identifier identifies the frame. The identified frame is returned to the requesting component.

Image frame abstraction model for image codecs

Systems for rendering Effect graphs for non-destructively processing digital image data which integrate Central Processing Unit (CPU) processing and Graphics Processing Unit (GPU) processing are provided. Additionally provided are systems for processing digital image data utilizing Effect graphs. The systems of the present invention integrate CPU processing and GPU processing to facilitate accelerated rendering of Effect graphs and, consequently, accelerated processing of digital images. Methods for processing digital image data utilizing the systems herein described are also provided.

Non-destructive processing of digital image data

A method and architecture for reading and updating metadata. A policy component is arranged to receive a request to read or update metadata that may include metadata from a plurality of standards. Each metadata format potentially includes a field corresponding to the request. The policy component determines which fields to read or update in satisfying the request by consulting a repository. The repository includes mappings that map information included in the request (e.g., a path) to locations in the metadata corresponding to the request. The policy component uses the locations to read or update the metadata.

Abstracted metadata policy component and related architecture

A system and method for automatically grouping items, e.g., digital photographs, by analyzing the statistical distributions of at least one piece of information associated with each of the items, for instance, creation time, are provided. In one embodiment, the creation time may be converted into a numeric value and the numeric values placed in ascending order such that gaps between adjacent numeric values may be determined. Subsequently, a median difference may be determined as well as an optimum difference. The items may then be organized into groups such that an additional group is created each time the difference between the numeric value for an item and the numeric value for the adjacent item located therebelow in the ascending array is greater than the optimum difference.

System and method for automatically grouping items

Described is the adding of padding to a content file (e.g., image and/or audiovisual file) for anticipated subsequent use such that the content file need not be re-encoded to add metadata and other data to it. For example, padding may be added to an image file directly after an image file directory, whereby the directory can grow as needed into the padded space. Metadata can likewise be added to the padding, e.g., from the opposite direction that the directory extends into the padding. The padding may be added on the first re-encoding of the file, reducing or eliminating the need for subsequent re-encoding. The padding also may be used to track dead space in the file, with adjacent regions of dead space coalesced into a single region of dead space.

Padding management for content files

A unified method for processing Raw and pre-processed digital images is provided. The method may include receiving digital image data and determining if the data is Raw or pre-processed. If the digital image data is Raw, the method may further include receiving modifications to the Raw image data, applying a sensor profile and an output profile to the data to create output-ready digital image data, and outputting the output-ready digital image data. If, on the other hand, the digital image data is pre-processed, the method may further include receiving a sensor profile that has been applied to the pre-processed image data (or a proxy therefore), converting the pre-processed image data to linear image data using the sensor profile, receiving modifications to the linear image data, applying the sensor profile and an output profile to the data to create output-ready digital image data, and outputting the output-ready digital image data.

Growing community of inventors

Redmond, WA, United States of America

Tomasz S M Kasperkiewicz