Apple Patent Application Improves Mandarin Chinese Translation
|Written by Steve Brachmann
Twitter | LinkedIn
Posted: February 16, 2013 @ 11:09 am
This Thursday was a relatively quiet day for Apple published patent applications, as the U.S. Patent & Trademark Office only released 16 of the documents, which is a fair amount less than usual. Many of the applications listed here focus on improvements to media processing and storage, an area of computer systems which has long been a focus for Apple. Other patent applications include improvements to Mandarin Chinese language translations and methods of grading display screens for light leakage.
Method for Disambiguating Multiple Readings in Language Conversion
U.S. Patent Application No. 20130041647
Language conversion software has been available for years and is widely accessible through the Internet. Current models of language conversion between any two written languages relies on statistical models of probability that help the software identify probable word combinations based on language used in the corpus, or a large body of texts used as a basis for proper language use in one language.
This model, however, does not do a good job of discerning between heteronyms, words that are spelled the same but have different meanings or pronunciations. This is especially problematic for Pinyin transcriptions of Mandarin Chinese, a language in which written characters can have many different meanings or pronunciations. Apple hopes that the system of language conversion it’s seeking to protect in this application will allay this problem and better discern the proper word choice between heteronyms.
As claim 1 describes, Apple hopes to protect:
“A method, comprising: receiving an input data to be converted into a set of characters comprising a symbolic representation of the input data in a target symbolic system; and using a language model that distinguishes between a first reading and a second reading of a character of the target symbolic system to determine a probability that the heteronymous character should be used to represent a corresponding portion of the input data.”
Categorization of Digital Media Based on Media Characteristics
U.S. Patent Application No. 20130041903
Tags are incredibly important when trying to index digital media, like videos or photos. Tags are words, phrases or other bits of data that act as keywords, improving the ability for browsers to find those items while searching for relevant terms. Tags can be associated with any piece of content uploaded to digital media servers by a computer, laptop or a handheld device, like a mobile phone.
Apple’s new tagging system, outlined in this patent application, improves the categorization of content by associating relevant metadata tags, including the identities of those appearing in photos or videos or the geographical location from which each photo was taken. This can help improve the organization of digital media files during a user search.
Claim 2 (Claim 1 cancelled) of this Apple patent application seeks protection for:
“A non-transitory program storage device, readable by a processor and comprising instructions stored thereon to cause one or more processors to: receive a first media object comprising a first digital image, a first feature, and a first set of metadata, wherein the first feature includes at least a portion of a first person’s face, and wherein the first set of metadata includes a first metadata tag identifying the first person; receive a second media object comprising a second digital image and a second set of metadata, wherein the second set of metadata does not comprise a metadata tag identifying the second person when the second media object is received; detect a second feature of the second digital image, wherein the second feature includes at least a portion of a second person’s face; identify one or more similarities between the first feature and the detected second feature; determine, based on the one or more identified similarities, that the first feature and detected second feature comprise portions of the same person’s face; and modify the second set of metadata to include a second metadata tag identifying the first person.”
Method and Apparatus for Managing Image-Processing Operations
U.S. Patent Application No. 20130039601
Image processing technologies available on computer systems continue to become more sophisticated with time. However, different technologies can be time-consuming as they don’t share a common interface, requiring users to interact with multiple programs to finish a single project. Extracting a still image from video code and converting it into a useful file format, for example, may require a user to open two or more programs, draining processor resources.
The system of image processing laid out in this Apple patent application utilizes pixel buffers to aid in rendering an image from one program interface onto a different interface. The pixel buffers maintain the image’s quality by retaining the cropping, pixel formatting and color from the original image representation. As Claim 1 of this patent application describes, Apple is seeking protection for:
“A non-transitory computer-readable storage medium containing instructions that when executed cause a computer to perform a method for image conversion, the method comprising: receiving a request from a client for one or more pixel buffers containing a converted version of an image representation; determining if the one or more pixel buffers can be obtained from a component that generates pixel buffers based, at least in part, on a predetermined source image representation type; obtaining the one or more pixel buffers from the component if it is determined that the component can provide the one or more pixel buffers; and if it is determined that the component cannot provide the one or more pixel buffers, then causing the component to: generate one or more intermediate pixel buffers; generate, by a second component, a sequence of converters to convert ert the one or more intermediate pixel buffers; and call the sequence of converters to generate the one or more pixel buffers.”
Display Light Leakage
U.S. Patent Application No. 20130038740
When a new display screen product is being marketed for electronic devices or computers, there are a number of specifications that are touted as characteristics of the new technology. These usually include the screen’s refresh rate, resolution and contrast capability, among others. Light leakage is another characteristic of display screens that can greatly affect user satisfaction. However, there exist no quality methods of testing and grading a display screen for light leakage during the manufacturing process.
Apple has created a new method of grading the light leakage of electronic displays that it is seeking to protect in this intellectual property patent application. The system includes a method of capturing images of the display screen that can be judged against a threshold parameter for light leakage. An image model of the screen is also rendered into a set of blocks that creates an intensity map to further identify leakage.
Claim 1 of this patent application seeks protection for:
“A method for rating display quality comprising: capturing an image of the display; computing a first parameter value from the captured image; computing a second parameter value from the captured image; comparing the first parameter value against a threshold; if the first parameter value exceeds the threshold, computing a grade wherein the second parameter is given a first weighting; if the first parameter value is below the threshold, computing a grade wherein the second parameter is given a second weighting which is greater than the first weighting.”
Low-Power GPU States for Reducing Power Consumption
U.S. Patent Application No. 20130038615
Improvements to high quality 2D and 3D graphics, as well as animated images, demand even more resources from a computer’s processing system than ever before. As digital graphic quality continued developing, computer and electronics manufacturers began creating more powerful graphics-processing units (GPUs) to process these images for display rendering. Although current GPUs are capable of processing very complex images, they create a major drain on both battery and processor power.
This system, which Apple’s patent application introduces, can create an efficient response to resource use on computers running both a high-power GPU and a low-power GPU. Typically, the display functions of a computer system, which render the computer interface on the display monitor, are always processed through the computer’s main GPU, typically the higher-powered unit. This system would monitor the system to identify an idle state, triggering a request to run the computer’s display through the low-power GPU unit until the computer wakes up.
Claim 1 of this Apple patent application would protect:
“A method for driving a display from a computer system, comprising: detecting a disabling condition in a first graphics-processing unit (GPU) used to drive the display, wherein the disabling condition causes the first GPU to be disabled; and upon detecting the disabling condition: switching from using the first GPU to using a second GPU to drive the display; and placing the first GPU into a low-power state, wherein the low-power state reduces a power consumption of the computer system.”