Generating preference indices for image content

Patent metadata

Field	Value
Publication number	US-2017193336-A1
Application number	US-201715463774-A
Country	US
Kind code	A1
Filing date	Mar 20, 2017
Priority date	Dec 22, 2014
Publication date	Jul 6, 2017
Grant date	—

Abstract

Briefly, embodiments of methods and/or systems of generating preference indices for contiguous portions of digital images are disclosed. For one embodiment, as an example, parameters of a neural network may be developed to generate object labels for digital images. The developed parameters may be transferred to a neural network utilized to generate signal sample value levels corresponding to preference indices for contiguous portions of digital images.

First claim

1 . A method of generating preference indices for images utilizing special-purpose computing devices to operate as neural networks, in which the special-purpose computing devices include one or more processors and one or more memory devices, comprising:
accessing computer instructions from the one or more memory devices of the special-purpose computing devices for execution on the one or more processors of the special-purpose computing devices; executing the accessed computer instructions utilizing the special-purpose computing devices; and storing, in the one or more memory devices of the special-purpose computing devices, results of having executed the accessed computer instructions on the one or more processors of the special-purpose computing devices, wherein the computer instructions comprise instructions for generating a term-independent preference index for a contiguous portion of an image using at least a second neural network of the neural networks, the second neural network including parameters transferred from a first neural network of the neural networks to identify one or more object labels for a captured image via at least one transfer learning classifier to obtain signal sample vectors from an input signal side of a fully-connected layer of a second neural network.

2 . The method of claim 1 , wherein the term-independent preference index comprises a sentiment index.

3 . The method of claim 1 , wherein the term-independent preference index comprises an odd number of allowed value levels.

4 . The method of claim 1 , wherein the second neural network comprises a convolutional neural network.

5 . The method of claim 4 , wherein the convolutional neural network comprises a deep convolutional neural network.

6 . The method of claim 1 , wherein the first neural network comprises a plurality of convolutional layers and a plurality of fully-connected layers.

7 . The method of claim 1 , wherein the generating the term-independent preference index includes generating the term-independent preference index via classification of signal samples generated by one or more convolutional layers of the second neural network.

8 . The method of claim 7 , wherein the classification comprises classification via a machine learning process.

9 . The method of claim 8 , wherein the machine learning process comprises a support vector machine process.

10 . The method of claim 7 , wherein the classification comprises classification via a regression process.

11 . An apparatus, comprising:
special-purpose computing devices to operate as neural networks, the special-purpose computing devices including one or more processors and one or more memory devices, to execute computer instructions accessed from the one or more memory devices, the special-purpose computing devices to store in the one or more memory devices results to be generated from the execution of the computer instructions on the one or more processors; the computer instructions to be executed comprising instructions for generating preference indices for image content, wherein
the computer instructions further to comprise instructions to:
generate a term-independent preference index for a contiguous portion of an image via at least a second neural network of the neural networks, the second neural network to access parameters to be transferred from a first neural network of the neural networks to identify one or more object labels for a captured image via at least one transfer learning classifier to obtain signal sample vectors from an input side of a fully-connected layer of the second neural network.

12 . The apparatus of claim 11 , the term-independent preference index is to comprise a sentiment index.

13 . The apparatus of claim 11 , wherein the term-independent preference index is to comprise an odd number of allowed value levels.

14 . The apparatus of claim 11 , wherein the second neural network is to comprise a deep convolutional neural network.

15 . The apparatus of claim 11 , wherein the term-independent preference index is to be generated via classification of signal samples generated by one or more convolutional layers of the second neural network.

16 . The apparatus of claim 11 , wherein the term-independent preference index is to be generated via classification of signal samples generated by one or more convolutional layers of the second neural network via a machine learning process.

17 . An apparatus to generate a term-independent preference index for images utilizing special-purpose computing devices configured as neural networks including one or more processors and one or more memory devices, comprising;
means for accessing computer instructions for execution on the one or more processors of the special-purpose computing devices; means for executing the accessed computer instructions utilizing the special-purpose computing devices; means for storing, in the one or more memory devices of the special-purpose computing devices, results of having executed the accessed computer instructions on the one or more processors of the special-purpose computing devices; and means for utilizing parameters developed for a first neural network of the neural networks, to identify one or more object labels for a captured image, via at least one transfer learning classifier to obtain signal sample vectors from an input side of a fully-connected layer of a second neural network of the neural networks.

18 . The apparatus of claim 17 , further comprising means for classifying signal samples generated by one or more convolutional layers of the neural network.

19 . The apparatus of claim 18 , wherein the means for classifying the signal samples comprises means for classifying via a machine learning process.

20 . The apparatus of claim 17 , further comprising means for generating an odd number of allowed levels of the term-independent preference index.

Assignees

• Yahoo Inc

Inventors

• Cetintas Suleyman
• Lee Kuang-Chih
• Li Jia

Classifications

• G06V10/765

  using rules for classification or partitioning the feature space · CPC title

• G06V10/75

  Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06F18/24

  Classification techniques · CPC title

• G06V30/413 Primary

  Classification of content, e.g. text, photographs or tables · CPC title