Adobe, Inc.

Nokia Corporation

Nokia Technologies Oy

Xiaohui Shen

Zhe Lin

Jianming Zhang

Radomir Mech

Jonathan Worthen Brandt

Xin Lu

Kalyan Krishna Sunkavalli

Jimei Yang

Brian Lynn Price

Haoxiang Li

Scott David Cohen

Chen Fang

Jiahui Yu

Yufei Wang

Mingyang Ling

Yi-Hsuan Tsai

Jen-Chan Jeff Chien

Joon-Young Lee

Jonathan Eisenmann

Gavin Stuart Peter Miller

Anant Gilra

Mehmet Yumer

Emiliano Gambaretto

Jason Wen Yong Kuen

Peng Wang

Zijun Wei

Marc-André Gardner

Chenxi Liu

Shu Kong

Jian Ren

Hailin Jin

Zhaowen Wang

Oliver Wang

Yilin Wang

Sarah Kong

Zhihong Ding

Bryan Russell

Chih-Yao Hsieh

I-Ming Pao

Michael Kaplan

Lance Williams

Brett Michael Butterfield

Lijun Wang

Yingwei Li

Jason Kuen

Yingcong Chen

Hengshuang Zhao

Yibing Song

Jae-Pil Heo

Gang Hua

Jianchao Yang

Ning Yu

Shanghang Zhang

Wei-Chih Hung

Methods and systems are provided for generating enhanced image. A neural network system is trained where the training includes training a first neural network that generates enhanced images conditioned on content of an image undergoing enhancement and training a second neural network that designates realism of the enhanced images generated by the first neural network. The neural network system is trained by determine loss and accordingly adjusting the appropriate neural network(s). The trained neural network system is used to generate an enhanced aesthetic image from a selected image where the output enhanced aesthetic image has increased aesthetics when compared to the selected image.

Aesthetics-guided image enhancement

Approaches are described for determining facial landmarks in images. An input image is provided to at least one trained neural network that determines a face region (e.g., bounding box of a face) of the input image and initial facial landmark locations corresponding to the face region. The initial facial landmark locations are provided to a 3D face mapper that maps the initial facial landmark locations to a 3D face model. A set of facial landmark locations are determined from the 3D face model. The set of facial landmark locations are provided to a landmark location adjuster that adjusts positions of the set of facial landmark locations based on the input image. The input image is presented on a user device using the adjusted set of facial landmark locations.

Neural network based face detection and landmark localization

In implementations of object detection in images, object detectors are trained using heterogeneous training datasets. A first training dataset is used to train an image tagging network to determine an attention map of an input image for a target concept. A second training dataset is used to train a conditional detection network that accepts as conditional inputs the attention map and a word embedding of the target concept. Despite the conditional detection network being trained with a training dataset having a small number of seen classes (e.g., classes in a training dataset), it generalizes to novel, unseen classes by concept conditioning, since the target concept propagates through the conditional detection network via the conditional inputs, thus influencing classification and region proposal. Hence, classes of objects that can be detected are expanded, without the need to scale training databases to include additional classes.

Object detection in images

Embodiments of the present invention provide an automated image tagging system that can predict a set of tags, along with relevance scores, that can be used for keyword-based image retrieval, image tag proposal, and image tag auto-completion based on user input. Initially, during training, a clustering technique is utilized to reduce cluster imbalance in the data that is input into a convolutional neural network (CNN) for training feature data. In embodiments, the clustering technique can also be utilized to compute data point similarity that can be utilized for tag propagation (to tag untagged images). During testing, a diversity based voting framework is utilized to overcome user tagging biases. In some embodiments, bigram re-weighting can down-weight a keyword that is likely to be part of a bigram based on a predicted tag set.

Accurate tag relevance prediction for image search

Embodiments of the present invention are directed to facilitating region of interest preservation. In accordance with some embodiments of the present invention, a region of interest preservation score using adaptive margins is determined. The region of interest preservation score indicates an extent to which at least one region of interest is preserved in a candidate image crop associated with an image. A region of interest positioning score is determined that indicates an extent to which a position of the at least one region of interest is preserved in the candidate image crop associated with the image. The region of interest preservation score and/or the preserving score are used to select a set of one or more candidate image crops as image crop suggestions.

Preserving regions of interest in automatic image cropping

The present disclosure relates to training and utilizing an image exposure transformation network to generate a long-exposure image from a single short-exposure image (e.g., still image). In various embodiments, the image exposure transformation network is trained using adversarial learning, long-exposure ground truth images, and a multi-term loss function. In some embodiments, the image exposure transformation network includes an optical flow prediction network and/or an appearance guided attention network. Trained embodiments of the image exposure transformation network generate realistic long-exposure images from single short-exposure images without additional information.

Utilizing an image exposure transformation neural network to generate a long-exposure image from a single short-exposure image

Semantic segmentation techniques and systems are described that overcome the challenges of limited availability of training data to describe the potentially millions of tags that may be used to describe semantic classes in digital images. In one example, the techniques are configured to train neural networks to leverage different types of training datasets using sequential neural networks and use of vector representations to represent the different semantic classes.

Semantic class localization digital environment

Systems and techniques for estimating illumination from a single image are provided. An example system may include a neural network. The neural network may include an encoder that is configured to encode an input image into an intermediate representation. The neural network may also include an intensity decoder that is configured to decode the intermediate representation into an output light intensity map. An example intensity decoder is generated by a multi-phase training process that includes a first phase to train a light mask decoder using a set of low dynamic range images and a second phase to adjust parameters of the light mask decoder using a set of high dynamic range image to generate the intensity decoder.

High dynamic range illumination estimation

Predicting patch displacement maps using a neural network is described. Initially, a digital image on which an image editing operation is to be performed is provided as input to a patch matcher having an offset prediction neural network. From this image and based on the image editing operation for which this network is trained, the offset prediction neural network generates an offset prediction formed as a displacement map, which has offset vectors that represent a displacement of pixels of the digital image to different locations for performing the image editing operation. Pixel values of the digital image are copied to the image pixels affected by the operation.

Predicting patch displacement maps using a neural network

Various embodiments describe frame selection based on training and using a neural network. In an example, the neural network is a convolutional neural network trained with training pairs. Each training pair includes two training frames from a frame collection. The loss function relies on the estimated quality difference between the two training frames. Further, the definition of the loss function varies based on the actual quality difference between these two frames. In a further example, the neural network is trained by incorporating facial heatmaps generated from the training frames and facial quality scores of faces detected in the training frames. In addition, the training involves using a feature mean that represents an average of the features of the training frames belonging to the same frame collection. Once the neural network is trained, a frame collection is input thereto and a frame is selected based on generated quality scores.

Frame selection based on a trained neural network

Digital image completion by learning generation and patch matching jointly is described. Initially, a digital image having at least one hole is received. This holey digital image is provided as input to an image completer formed with a dual-stage framework that combines a coarse image neural network and an image refinement network. The coarse image neural network generates a coarse prediction of imagery for filling the holes of the holey digital image. The image refinement network receives the coarse prediction as input, refines the coarse prediction, and outputs a filled digital image having refined imagery that fills these holes. The image refinement network generates refined imagery using a patch matching technique, which includes leveraging information corresponding to patches of known pixels for filtering patches generated based on the coarse prediction. Based on this, the image completer outputs the filled digital image with the refined imagery.

Digital image completion by learning generation and patch matching jointly

Compositing aware digital image search techniques and systems are described that leverage machine learning. In one example, a compositing aware image search system employs a two-stream convolutional neural network (CNN) to jointly learn feature embeddings from foreground digital images that capture a foreground object and background digital images that capture a background scene. In order to train models of the convolutional neural networks, triplets of training digital images are used. Each triplet may include a positive foreground digital image and a positive background digital image taken from the same digital image. The triplet also contains a negative foreground or background digital image that is dissimilar to the positive foreground or background digital image that is also included as part of the triplet.

Compositing aware digital image search

Digital image completion using deep learning is described. Initially, a digital image having at least one hole is received. This holey digital image is provided as input to an image completer formed with a framework that combines generative and discriminative neural networks based on learning architecture of the generative adversarial networks. From the holey digital image, the generative neural network generates a filled digital image having hole-filling content in place of holes. The discriminative neural networks detect whether the filled digital image and the hole-filling digital content correspond to or include computer-generated content or are photo-realistic. The generating and detecting are iteratively continued until the discriminative neural networks fail to detect computer-generated content for the filled digital image and hole-filling content or until detection surpasses a threshold difficulty. Responsive to this, the image completer outputs the filled digital image with hole-filling content in place of the holey digital image's holes.

Digital image completion using deep learning

The present disclosure relates to systems, methods, and non-transitory computer readable media for utilizing a deep neural network-based model to identify similar digital images for query digital images. For example, the disclosed systems utilize a deep neural network-based model to analyze query digital images to generate deep neural network-based representations of the query digital images. In addition, the disclosed systems can generate results of visually-similar digital images for the query digital images based on comparing the deep neural network-based representations with representations of candidate digital images. Furthermore, the disclosed systems can identify visually similar digital images based on user-defined attributes and image masks to emphasize specific attributes or portions of query digital images.

Identifying visually similar digital images utilizing deep learning

Growing community of inventors

San Jose, CA, United States of America

Xiaohui Shen