IP Library › Granted Patent US 12,608,596
Granted Patent B2
US 12,608,596 · App. 18/270,685 · Granted Apr 21, 2026

Computer vision models using global and local information

Inventors: Ruofei Du (San Francisco, CA); Yinda Zhang (Daly City, CA); Weihao Zeng (Taichung city, TW)
Assignee: Google LLC
G06N3/0464G06N3/084G06V10/42G06V10/44G06V10/82
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,608,596
App. No.
18/270,685
Granted
Apr 21, 2026
Kind
B2
Abstract

A system including a computer vision model configured to perform a machine learning task is described. The computer vision model includes multiple wrapped convolutional layers, in which each wrapped convolutional layer includes a respective convolutional layer configured to receive, for each time step of multiple time steps, a layer input and to process the layer input to generate an initial output for the current time step, and a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, the feature vector representing local information of the wrapped convolutional layer. The model includes a summarization module configured to receive the feature vectors and to process the feature vectors to generate a revision vector for the current time step, the revision vector representing global information of the plurality of wrapped convolutional layers.

Claims (96)

1 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement a neural network that is configured to perform a machine learning task by receiving an input image and processing the input image to generate a network output, the neural network comprising:

a plurality of wrapped convolutional neural network layers, wherein each wrapped convolutional neural network layer of the plurality of wrapped convolutional neural network layers comprises:

a respective convolutional neural network layer configured to receive, for each time step of a plurality of time steps, a layer input and to process the layer input to generate an initial output for the current time step, and

a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

a summarization module configured to receive the feature vectors generated by the respective note-taking modules for the current time step and to process the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

wherein each wrapped convolutional neural network layer further comprises a respective amendment module configured to, for each time step after a first time step in the plurality of time steps:

receive an amendment module input for the current time step, the amendment module input comprising one or more of: (i) a previous revision vector generated by the summarization module for a previous time step, (ii) a previous feature vector generated by the respective note-taking module of the wrapped convolutional neural network layer for the previous time step, or iii) the feature vector generated by the respective note-taking module of the wrapped convolutional neural network layer for the current time step,

process the amendment module input to generate a set of amendment parameters, and

receive the initial output for the current time step and process the initial output based on the set of amendment parameters to generate a layer output for the wrapped convolutional neural network layer.

2 . The system of claim 1 , wherein the set of amendment parameters comprises, for each channel of a plurality of channels of the initial output, a respective pair of control variables, and

wherein the respective amendment module is configured to process the initial output to generate the layer output by:

scaling and shifting a variance and bias of each channel of the plurality of channels of the initial output using the respective pair of control variables of the channel.

3 . The system of claim 2 , wherein the respective amendment module is further configured to apply instance normalization to the initial output before scaling and shifting the variance and bias of each channel of the plurality of channels.

4 . The system of claim 1 , wherein the feature vector is a one-dimensional (1D) vector.

5 . The system of claim 1 , wherein the neural network is a convolutional neural network.

6 . The system of claim 1 , wherein the layer input is one of the following: (i) the input image, (ii) an input derived from the input image, (iii) an output of a preceding wrapped convolutional neural network layer in the plurality of wrapped convolutional neural network layers, or (iv) an output of another component of the neural network.

7 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement a neural network that is configured to perform a machine learning task by receiving an input image and processing the input image to generate a network output, the neural network comprising:

a plurality of wrapped convolutional neural network layers, wherein each wrapped convolutional neural network layer of the plurality of wrapped convolutional neural network layers comprises:

a respective convolutional neural network layer configured to receive, for each time step of a plurality of time steps, a layer input and to process the layer input to generate an initial output for the current time step, and

a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

a summarization module configured to receive the feature vectors generated by the respective note-taking modules for the current time step and to process the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

wherein the respective note-taking module comprises a pre-processing layer, a pooling layer, and a multi-layer perceptron (MLP) layer, and wherein the respective note-taking module is configured to process the initial output to generate the feature vector for the current time step by:

applying, using the pre-processing layer, a pre-processing operation to the initial output to generate a first intermediate output,

applying, using the pooling layer, a pooling operation to the first intermediate output to generate a second intermediate output,

flattening the second intermediate output to generate a third intermediate output, and

processing, using the MLP layer, the third intermediate output to generate the feature vector for the current time step.

8 . The system of claim 7 , wherein the pre-processing layer comprises a 1×1 convolutional neural network layer, and

wherein applying the pre-processing operation to the initial output comprises applying a 1×1 convolution to the initial output to reduce a channel dimension of the initial output.

9 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement a neural network that is configured to perform a machine learning task by receiving an input image and processing the input image to generate a network output, the neural network comprising:

a plurality of wrapped convolutional neural network layers, wherein each wrapped convolutional neural network layer of the plurality of wrapped convolutional neural network layers comprises:

a respective convolutional neural network layer configured to receive, for each time step of a plurality of time steps, a layer input and to process the layer input to generate an initial output for the current time step, and

a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

a summarization module configured to receive the feature vectors generated by the respective note-taking modules for the current time step and to process the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

wherein the respective note-taking module comprises a pooling layer and a multi-layer perceptron (MLP) layer, and wherein the respective note-taking module is configured to process the initial output to generate the feature vector for the current time step by:

applying, using a pooling layer, a pooling operation to the initial output to generate a second intermediate output,

flattening the second intermediate output to generate a third intermediate output, and

processing, using the MLP layer, the third intermediate output to generate the feature vector for the current time step.

10 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement a neural network that is configured to perform a machine learning task by receiving an input image and processing the input image to generate a network output, the neural network comprising:

a plurality of wrapped convolutional neural network layers, wherein each wrapped convolutional neural network layer of the plurality of wrapped convolutional neural network layers comprises:

a respective convolutional neural network layer configured to receive, for each time step of a plurality of time steps, a layer input and to process the layer input to generate an initial output for the current time step, and

a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

a summarization module configured to receive the feature vectors generated by the respective note-taking modules for the current time step and to process the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

wherein the summarization module comprises a second multi-layer perception (MLP) layer and is configured to process the feature vectors to generate a revision vector for the current time step by:

concatenating the feature vectors generated by the respective note-taking modules for the current time step to generate a concatenated feature vector, and

processing the concatenated feature vector using the second MLP layer to generate the revision vector for the current time step.

11 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement a neural network that is configured to perform a machine learning task by receiving an input image and processing the input image to generate a network output, the neural network comprising:

a plurality of wrapped convolutional neural network layers, wherein each wrapped convolutional neural network layer of the plurality of wrapped convolutional neural network layers comprises:

a respective convolutional neural network layer configured to receive, for each time step of a plurality of time steps, a layer input and to process the layer input to generate an initial output for the current time step, and

a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

a summarization module configured to receive the feature vectors generated by the respective note-taking modules for the current time step and to process the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

wherein the initial output is a three-dimensional (3D) vector.

12 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement a neural network that is configured to perform a machine learning task by receiving an input image and processing the input image to generate a network output, the neural network comprising:

a plurality of wrapped convolutional neural network layers, wherein each wrapped convolutional neural network layer of the plurality of wrapped convolutional neural network layers comprises:

a respective convolutional neural network layer configured to receive, for each time step of a plurality of time steps, a layer input and to process the layer input to generate an initial output for the current time step, and

a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

a summarization module configured to receive the feature vectors generated by the respective note-taking modules for the current time step and to process the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

wherein the revision vector is a one-dimensional (1D) vector.

13 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement a neural network that is configured to perform a machine learning task by receiving an input image and processing the input image to generate a network output, the neural network comprising:

a plurality of wrapped convolutional neural network layers, wherein each wrapped convolutional neural network layer of the plurality of wrapped convolutional neural network layers comprises:

a respective convolutional neural network layer configured to receive, for each time step of a plurality of time steps, a layer input and to process the layer input to generate an initial output for the current time step, and

a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

a summarization module configured to receive the feature vectors generated by the respective note-taking modules for the current time step and to process the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

wherein the neural network is a Generative Adversarial Network (GAN).

14 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to implement a neural network that is configured to perform a machine learning task by receiving an input image and processing the input image to generate a network output, the neural network comprising:

a plurality of wrapped convolutional neural network layers, wherein each wrapped convolutional neural network layer of the plurality of wrapped convolutional neural network layers comprises:

a respective convolutional neural network layer configured to receive, for each time step of a plurality of time steps, a layer input and to process the layer input to generate an initial output for the current time step, and

a respective note-taking module configured to receive the initial output and to process the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

a summarization module configured to receive the feature vectors generated by the respective note-taking modules for the current time step and to process the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

the machine learning task is one of the following: a super resolution task, a depth estimation task, a style transfer task, an image classification task, an object recognition task, an optical character recognition (OCR) task, or a semantic segmentation task.

15 . A computer-implemented method comprising:

performing the following operations at each of a plurality of time steps:

for each of a plurality of wrapped convolutional neural network layers of a neural network:

receiving, by the wrapped convolutional neural network layer, a layer input and processing the layer input to generate an initial output for the current time step, and

receiving, by the wrapped convolutional neural network layer, the initial output and processing the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

receiving, by a summarization module of the neural network, the feature vectors generated for the current time step and processing the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

further comprising:

at each time step after a first time step in the plurality of time steps, performing the following operations for each of the plurality of wrapped convolutional neural network layers:

receiving an amendment module input for the current time step, the amendment module input comprising one or more of: (i) a previous revision vector generated by the summarization module for a previous time step, (ii) a previous feature vector generated by the wrapped convolutional neural network layer for the previous time step, or (iii) the feature vector generated by the wrapped convolutional neural network layer for the current time step,

processing the amendment module input to generate a set of amendment parameters, and

processing the initial output for the current time step based on the set of amendment parameters to generate a layer output for the wrapped convolutional neural network layer.

16 . A computer-implemented method comprising:

performing the following operations at each of a plurality of time steps:

for each of a plurality of wrapped convolutional neural network layers of a neural network:

receiving, by the wrapped convolutional neural network layer, a layer input and processing the layer input to generate an initial output for the current time step, and

receiving, by the wrapped convolutional neural network layer, the initial output and processing the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

receiving, by a summarization module of the neural network, the feature vectors generated for the current time step and processing the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

wherein the layer input is one of the following: (i) an input image of the neural network, (ii) an input derived from the input image, (iii) a layer output of a preceding wrapped convolutional neural network layer in the plurality of wrapped convolutional neural network layers, or (iv) an output of another component of the neural network.

17 . One or more non-transitory computer storage media storing instructions that, when executed by one or more computers, cause the one or more computers to perform the following operations at each of a plurality of time steps:

for each of a plurality of wrapped convolutional neural network layers of a neural network:

receiving, by the wrapped convolutional neural network layer, a layer input and processing the layer input to generate an initial output for the current time step, and

receiving, by the wrapped convolutional neural network layer, the initial output and processing the initial output to generate a feature vector for the current time step, wherein the feature vector represents local information of the wrapped convolutional neural network layer; and

receiving, by a summarization module of the neural network, the feature vectors generated for the current time step and processing the feature vectors to generate a revision vector for the current time step, wherein the revision vector represents global information of the plurality of wrapped convolutional neural network layers;

at each time step after the first time step in the plurality of time steps, performing the following operations for each of the plurality of wrapped convolutional neural network layers:

receiving an amendment module input for the current time step, the amendment module input comprising one or more of: (i) a previous revision vector generated by the summarization module for a previous time step, (ii) a previous feature vector generated by the wrapped convolutional neural network layer for the previous time step, or (iii) the feature vector generated by the wrapped convolutional neural network layer for the current time step,

processing the amendment module input to generate a set of amendment parameters, and

processing the initial output for the current time step based on the set of amendment parameters to generate a layer output for the wrapped convolutional neural network layer.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 10, 2023
From: DU, RUOFEI; ZHANG, YINDA; ZENG, WEIHAO
To: GOOGLE LLC
Reel/Frame 065170/0388 →
Continuity (1)
Related Publication 20240062046A1 · Feb 22, 2024
References Cited (77)
US 9613450B2 · Wang · 2017 [cited by examiner]
US 9730643B2 · Georgescu · 2017 [cited by examiner]
US 10424087B2 · Risser · 2019 [cited by examiner]
US 10635951B1 · Liu · 2020 [cited by examiner]
US 11216656B1 · Zia · 2022 [cited by examiner]
US 11651192B2 · Gabriel · 2023 [cited by examiner]
US 20170147544A1 · Modani · 2017 [cited by examiner]
US 20180075343A1 · van den Oord · 2018 [cited by examiner]
US 20180137417A1 · Theodorakopoulos · 2018 [cited by examiner]
US 20180225519A1 · Chen · 2018 [cited by examiner]
US 20180268284A1 · Ren · 2018 [cited by examiner]
US 20190171926A1 · Chen · 2019 [cited by examiner]
US 20190258671A1 · Bou · 2019 [cited by examiner]
US 20190327103A1 · Niekrasz · 2019 [cited by examiner]
US 20200380027A1 · Aggarwal · 2020 [cited by examiner]
US 20210201045A1 · Iyer · 2021 [cited by examiner]
US 20210201129A1 · Schmude · 2021 [cited by examiner]
US 20210232943A1 · Abishek Kumar · 2021 [cited by examiner]
US 20210303796A1 · Elsahar · 2021 [cited by examiner]
US 20220012365A1 · Garg · 2022 [cited by examiner]
US 20220300819A1 · Ren · 2022 [cited by examiner]
US 20230376677A1 · Choubey · 2023 [cited by examiner]
US 20240036932A1 · Croxford · 2024 [cited by examiner]
US 20240160837A1 · Agarwal · 2024 [cited by examiner]
US 20240202461A1 · Pagnoni · 2024 [cited by examiner]
US 20240242022A1 · Yee · 2024 [cited by examiner]
US 20250022107A1 · Qian · 2025 [cited by examiner]
WO WO2017106998A1 · 2017 [cited by examiner]
Aadil et al., “Improving super resolution methods via incremental residual learning,” Paper, Presented at 2019 IEEE International Conference on Image Processing, Taipei, Taiwan, Sep. 22-25, 2019, pp. 2836-2840. [cited by applicant]
Agustsson et al., “NTIRE 2017 challenge on single image super-resolution: Dataset and study,” Paper, Presented at the IEEE Conference on Computer Vision and Pattern Recognition Workshops, Honolulu HI, Jul. 21-26, 2017, … [cited by applicant]
Bello et al., “Attention augmented convolutional networks,” Paper, Presented at 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seoul, South Korea, Oct. 27-Nov. 2, 2019, pp. 3286-3295. [cited by applicant]
Bevilacqua et al., “Low-complexity single-image super-resolution based on nonnegative neighbor embedding,” Paper, Presented at Proceedings of the British Machine Vision Conference, Surrey, United Kingdom, Sep. 3-7, 2012… [cited by applicant]
Cabon et al., “Virtual KITTI 2,” CoRR, Submitted on Jan. 29, 2020, arXiv:2001.10773v1, 11 pages. [cited by applicant]
Carreira et al., “Human Pose Estimation with Iterative Error Feedback,” Paper, Presented at Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, Jun. 27-30, 2016, pp. 4733-4742. [cited by applicant]
Chen et al., “Dynamic convolution: attention over convolution kernels,” Paper, Presented at 2020 IEEE Conference on Computer Vision and Pattern Recognition, Seattle, WA, Jun. 13-19, 2020, pp. 11030-11039. [cited by applicant]
Chen et al., “Facilitating English-language reading performance by a digital reading annotation system with self-regulated learning mechanisms,” Journal of Educational Technology & Society, Jan. 2014, 17(1): 102-114. [cited by applicant]
Cho et al., “Learning phrase representations using RNN encoder-decoder for statistical machine translation,” CoRR, Submitted on Sep. 3, 2014, arXiv:1406.1078v3, 15 pages. [cited by applicant]
Cho et al., “On the properties of neural machine translation: Encoder-decoder approaches,” CoRR, Submitted on Oct. 7, 2014, arXiv:1409.1259v2, 9 pages. [cited by applicant]
Eigen et al., “Depth map prediction from a single image using a multi-scale deep network,” Paper, Presented at 28th Conference on Neural Information Processing Systems, Montreal, Canada, Dec. 8-13, 2014; Advances in Neu… [cited by applicant]
Fu et al., “Look closer to see better: recurrent attention convolutional neural network for fine-grained image recognition,” Paper, Presented at 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu,… [cited by applicant]
Godard et al., “Digging into self-supervised monocular depth estimation,” Paper, Presented at Proceedings of the 2019 IEEE International Conference on Computer Vision, Seoul, South Korea, Oct. 27-Nov. 2, 2019, pp. 3828-… [cited by applicant]
Gu et al., “Blind super-resolution with iterative kernel correction,” Paper, Presented at Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, Jun. 15-20, 2019, pp. 1604-16… [cited by applicant]
Gu et al., “Recent advances in convolutional neural networks,” CoRR, Published May 2018, Submitted on Oct. 19, 2017, arXiv:1512.07108v6; Pattern Recognition, May 2018, 77(c):1-38. [cited by applicant]
Han et al., “Image super-resolution via dual-state recurrent networks,” Paper, Presented at Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, Jun. 18-23, 2018, pp. 1654-1… [cited by applicant]
He et al., “Deep residual learning for image recognition,” Paper, Presented at Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, Jun. 27-30, 2016, pp. 770-778. [cited by applicant]
Huang et al., “Arbitrary style transfer in real-time with adaptive instance normalization,” Paper, Presented at Proceedings of the 2017 IEEE International Conference on Computer Vision, Venice, Italy, Oct. 22-29, 2017, … [cited by applicant]
Huang et al., “Single image super-resolution from transformed self-exemplars,” Paper, Presented at Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, MA, Jun. 7-12, 2015, pp. 5197-520… [cited by applicant]
International Preliminary Report on Patentability in International Appln. No. PCT/US2021/025044, mailed Oct. 12, 2023, 10 pages. [cited by applicant]
International Search Report and Written Opinion in International Appln. No. PCT/US2021/025044, mailed Dec. 22, 2021, 16 pages. [cited by applicant]
Islam et al., “Gated feedback refinement network for dense image labeling,” Paper, Presented at Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu HI, Jul. 21-26, 2017, pp. 3751-3759. [cited by applicant]
Johnson et al., “Perceptual losses for real-time style transfer and super-resolution,” Paper, Presented at European Conference on Computer Vision, Amsterdam, Netherlands, Oct. 11-14, 2016; Computer Vision, Sep. 17, 2016… [cited by applicant]
Karras et al., “A style-based generator architecture for generative adversarial networks,” Paper, Presented at Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, Jun. 15-… [cited by applicant]
Karras et al., “Analyzing and improving the image quality of stylegan,” Paper, Presented at Proceedings of the 2020 IEEE Conference on Computer Vision and Pattern Recognition, Seattle, WA, Jun. 13-19, 2020, pp. 8110-811… [cited by applicant]
Kim et al., “Deeply-recursive convolutional network for image super-resolution,” Paper, Presented at Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, Jun. 27-30, 2016, p… [cited by applicant]
Kingma et al., “Adam: A method for stochastic optimization,” CoRR, Submitted on Dec. 22, 2014, arXiv:1412.6980v1, 9 pages. [cited by applicant]
Krizhevsky et al., “Imagenet classification with deep convolutional neural networks,” Communications of the ACM, May 2017, 60(6):84-90. [cited by applicant]
Li et al., “DeepIM: Deep iterative matching for 6d pose estimation,” Paper, Presented at Proceedings of the 2018 European Conference on Computer Vision, Munich, Germany, Sep. 8-14, 2018; Computer Vision, Sep. 8, 2018, 1… [cited by applicant]
Li et al., “Feedback network for image super-resolution,” Paper, Presented at Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, Jun. 15-20, 2019, pp. 3867-3876. [cited by applicant]
Li et al., “Gated multiple feedback network for image super-resolution,” CoRR, Submitted on Jul. 10, 2019, arXiv:1907.04253v2, 17 pages. [cited by applicant]
Lin et al., “Microsoft COCO: Common objects in context,” 2014 European Conference on Computer Vision, Zurich, Switzerland, Sep. 6-12, 2014; CoRR, Submitted on Feb. 21, 2015, arXiv:1405.0312v3, 15 pages. [cited by applicant]
Liu et al., “Iterative Network for Image Super Resolution,” IEEE Transactions on Multimedia, May 10, 2021, 24:2259-2272. [cited by applicant]
Long et al., “Fully convolutional networks for semantic segmentation,” Paper, Presented at Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition, Boston, MA, Jun. 7-12, 2015, pp. 3431-3440. [cited by applicant]
Martin et al., “A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics,” Paper, Presented at Proceedings Eighth IEEE International Conf… [cited by applicant]
Matsui et al., “Sketch-based manga retrieval using manga109 dataset,” Multimedia Tools and Applications, Nov. 9, 2016, 76(20):21811-21838. [cited by applicant]
Simonyan et al., “Very deep convolutional networks for large-scale image recognition,” CoRR, Submitted on Dec. 23, 2014, arXiv:1409.1556v5, 13 pages. [cited by applicant]
Slotte et al., “Review and process effects of spontaneous note-taking on text comprehension,” Contemporary Educational Psychology, 1999, 24(1):1-20. [cited by applicant]
Su et al., “Effects of annotations and homework on learning achievement: An empirical study of scratch programming pedagogy,” Journal of Educational Technology & Society, Feb. 2015, 18(4):331-343. [cited by applicant]
Sun et al., “Distilling with residual network for single image super resolution,” Paper, Presented at 2019 IEEE International Conference on Multimedia and Expo, Shanghai, China, Jul. 8-12, 2019, pp. 1180-1185. [cited by applicant]
Wang et al., “Anytime stereo image depth estimation on mobile devices,” Paper, Presented at 2019 International Conference on Robotics and Automation, Montreal, Canada, May 20-24, 2019, pp. 5893-5900. [cited by applicant]
Wang et al., “Deep learning for image super-resolution: A survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Mar. 23, 2020, 43(10):3365-3387. [cited by applicant]
Wang et al., “Densefusion: 6d object pose estimation by iterative dense fusion,” Paper, Presented at Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, Jun. 15-20, 2019, pp. 3… [cited by applicant]
Wei et al., “Network iterative learning for dynamic deep neural networks via morphism,” Rejected Paper, submitted for review for the 2018 International Conference on Learning Representations, Vancouver, Canada, Apr. 30 … [cited by applicant]
Yao et al., “Recurrent MVSNet for high-resolution multiview stereo depth inference,” Paper, Presented at Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, Jun. 15-20, 20… [cited by applicant]
Zamir et al., “Feedback networks,” Paper, Presented at Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, Jul. 21-26, 2017, pp. 1308-1317. [cited by applicant]
Zeyde et al., “On Single Image Scale-Up Using Sparse-Representations,” Paper, Presented at International Conference on Curves and Surfaces, Avignon, France, Jun. 24-30, 2010, pp. 711-730. [cited by applicant]
Zhang et al., “Canet: Class-agnostic segmentation networks with iterative refinement and attentive few-shot learning,” Paper, Presented at Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recogniti… [cited by applicant]
Office Action in European Appln. No. 21720628.3, mailed on Apr. 4, 2025, 9 pages. [cited by applicant]