Techniques are disclosed for image matting. In particular, embodiments decompose the matting problem of estimating foreground opacity into the targeted subproblems of estimating a background using a first trained neural network, estimating a foreground using a second neural network and the estimated background as one of the inputs into the second neural network, and estimating an alpha matte using a third neural network and the estimated background and foreground as two of the inputs into the third neural network. Such a decomposition is in contrast to traditional sampling-based matting approaches that estimated foreground and background color pairs together directly for each pixel. By decomposing the matting problem into subproblems that are easier for a neural network to learn compared to traditional data-driven techniques for image matting, embodiments disclosed herein can produce better opacity estimates than such data-driven techniques as well as sampling-based and affinity-based matting approaches.

There is provided an image generation method and an image synthesis method each of which can achieve an entire image having excellent image quality. An image generation method according to an embodiment of the present invention includes: arranging an image taking apparatus, a first polarizing plate, an object, and a second polarizing plate in the stated order; arranging a retardation plate between the first polarizing plate and the second polarizing plate; and monochromatizing a color of the second polarizing plate recognized by the image taking apparatus to a color complementary to that of the object through use of the retardation plate.

Disclosed is a broadcast providing method implemented at an electronic device including processing circuitry. The broadcast providing method includes receiving, by the processing circuitry, a broadcast image from a broadcast server, and generating, by the processing circuitry, a final image by synthesizing the broadcast image with a personalized background image using a chroma key, the personalized background image being personalized for a user of the electronic device.

This disclosure describes a method including determining one or more object clusters from a plurality of frames of the video content. At least one of the one or more object clusters is an object cluster with movement through the plurality of frames. The method includes extracting the determined one or more object clusters from the plurality of frames to generate a set of frames having extracted one or more object clusters and outputting the set of frames having extracted one or more object clusters. This disclosure describes a method including receiving the set of frames having the extracted one or more object clusters, rendering one or more of the set of frames in a live camera feed of a device, and generating video content based on the rendered one or more set of frames and a user interacting with the extracted one or more object clusters.

This disclosure describes a method including determining one or more object clusters from a plurality of frames of the video content. At least one of the one or more object clusters is an object cluster with movement through the plurality of frames. The method includes extracting the determined one or more object clusters from the plurality of frames to generate a set of frames having extracted one or more object clusters and outputting the set of frames having extracted one or more object clusters. This disclosure describes a method including receiving the set of frames having the extracted one or more object clusters, rendering one or more of the set of frames in a live camera feed of a device, and generating video content based on the rendered one or more set of frames and a user interacting with the extracted one or more object clusters.

Techniques are disclosed for image matting. In particular, embodiments decompose the matting problem of estimating foreground opacity into the targeted subproblems of estimating a background using a first trained neural network, estimating a foreground using a second neural network and the estimated background as one of the inputs into the second neural network, and estimating an alpha matte using a third neural network and the estimated background and foreground as two of the inputs into the third neural network. Such a decomposition is in contrast to traditional sampling-based matting approaches that estimated foreground and background color pairs together directly for each pixel. By decomposing the matting problem into subproblems that are easier for a neural network to learn compared to traditional data-driven techniques for image matting, embodiments disclosed herein can produce better opacity estimates than such data-driven techniques as well as sampling-based and affinity-based matting approaches.

Systems and methods for creating and distributing professional quality pictorial souvenirs giving the illusion that guests of a facility were imaged at other locations, including making initial arrangements with guests, showing selections of background scene images at video displays and flashing chroma key images interspersed with frames of the scene images, taking key guest images in synch with the flashed key images, extracting guest image content from the key guest images and merging into selecting scene images, showing preview merges images for guest selection and providing souvenir portfolios that include merged images to guests or designees after making financial arrangements, including for payment to third parties for copyright content included in the souvenirs and with advertisers for promotional material included in the souvenirs.

Systems and methods for creating and distributing professional quality pictorial souvenirs giving the illusion that guests of a facility were imaged at other locations, including making initial arrangements with guests, showing selections of background scene images at video displays and flashing chroma key images interspersed with frames of the scene images, taking key guest images in synch with the flashed key images, extracting guest image content from the key guest images and merging into selecting scene images, showing preview merges images for guest selection and providing souvenir portfolios that include merged images to guests or designees after making financial arrangements, including for payment to third parties for copyright content included in the souvenirs and with advertisers for promotional material included in the souvenirs.

The disclosure provides an electronic apparatus and a controlling method thereof. The electronic apparatus of the disclosure includes a sensor, a camera, and a processor configured to control the electronic apparatus to: obtain information regarding a predetermined object among a plurality of objects included in an imaging region of the camera through the sensor, identify an object region corresponding to the predetermined object from the imaging region of the camera based on the obtained information, and based on the imaging region being captured through the camera, control the camera to not capture the object region.

Disclosed is an electronic device and a method to generate alpha masks of video frames in a video comprising a plurality of video frames including a first video frame and a second video frame following the first video frame, where the method comprising; obtaining a first alpha mask of the first video frame; providing a first downscaled video frame, wherein the first downscaled video frame is a lower resolution version of the first video frame; providing a first downscaled alpha mask of the first alpha mask; estimating a first primary coefficient and a first secondary coefficient based on the first downscaled video frame and the first downscaled alpha mask; and generating a second alpha mask for the second video frame based on the first primary coefficient and the first secondary coefficient.