Nippon Telegraph and Telephone Corporation

Masaaki Matsumura

Seishi Takamura

Atsushi Shimizu

Hideaki Kimata

Hajime Noto

Hirohisa Jozawa

Kazuki Okami

Hiroshi Fujii

Yoshiyuki Yashima

Megumi Isogai

Shohei Matsuo

Yoshinori Kusachi

Akihiko Takashima

An object-specific keypoint separation apparatus includes: an inference execution unit configured to receive a captured image capturing an object as an input and use a pre-trained model that has been trained in order to output a plurality of first maps and a plurality of second maps generated from the input captured image to output the plurality of first maps and the plurality of second maps, the plurality of first maps storing a vector describing a connection relationship of a keypoint of the object only around the keypoint, and the plurality of second maps representing a heat map configured to have a peak at coordinates at which the keypoint of the object appears; a map correction unit configured to correct the plurality of second maps using the plurality of first maps and the plurality of second maps; an upsampling unit configured to upsample the plurality of first maps; and an object-specific keypoint separation unit configured to separate keypoints for each object based on the plurality of upsampled first maps and the plurality of corrected second maps.

Apparatus for separating feature points for each object, method for separating feature points for each object and computer program

A unit derives displacement amounts of feature points in an entire region of a face shape. A unit derives separated displacement amounts, each of which is a displacement amount of each feature point for every deformation pattern of a partial region of the face shape, based on the displacement amounts of the feature points in the entire region of the face shape and deformation degrees for every deformation pattern of the partial regions. A unit derives normalized and separated displacement amounts, which are obtained by normalizing the separated displacement amounts, for every deformation pattern of the partial regions. A unit generates weight values of the feature points in the entire region of the face shape based on the normalized and separated displacement amounts for every deformation pattern of the partial regions of the face shape and the deformation degrees for every deformation pattern of the partial regions.

Weighted mask generating apparatus, weighted mask generating method and program

A subject tracking device includes a first histogram generation unit configured to generate a first histogram representing an appearance frequency of an image feature quantity of a subject region, for each segment and for each auxiliary segment of the image feature quantity, a second histogram generation unit configured to generate, for each candidate region, a second histogram representing an appearance frequency of the image feature quantity of an image of a candidate region for each segment and for each auxiliary segment of the image feature quantity, a difference derivation unit configured to compare the first histogram with the second histogram for each segment and for each auxiliary segment, and derive, for each candidate region, a total difference value of the appearance frequencies of the image feature quantities of the first histogram and the second histogram; and a region selection unit configured to select the candidate region having a minimum total difference value from among the candidate regions. The auxiliary segment is defined to straddle boundaries between the segments.

Object tracking apparatus, object tracking method and program

A learning device includes: a time-series information generation unit that obtains a first image group including a plurality of successive time-series images including a reference image and generates first time-series information based on a difference between the reference image and each of the images in the first image group other than the reference image; and a first learning unit that performs machine learning using the reference image and the first time-series information, thereby obtaining a first learning result used for estimating depth information on a target image, which is an image to be processed, and silhouette information on a subject captured in the target image based on the target image and second time-series information generated from a second image group including a plurality of successive time-series images including the target image.

Learning apparatus, estimation apparatus, learning method, estimation method, and computer programs

Setting a reference pixel signal from a decoded pixel signal in intra-prediction in which a prediction signal is generated in the same frame, acquiring prediction mode identification information that identifies an intra-prediction mode, generating a prediction signal based on the reference pixel signal and the prediction mode identification information, making a determination regarding correction of the prediction signal using a reference pixel having a shorter distance from a pixel that is a prediction target in the intra-prediction mode identified by the prediction mode identification information, and correcting the generated prediction signal according to a result of the determination are included in order to provide an image encoding method and an image decoding method capable of improving subjective image quality and coding efficiency using a closer decoded pixel in an intra-prediction process.

Intra-prediction encoding method, intra-prediction decoding method, intra-prediction encoding apparatus, intra-prediction decoding apparatus, program therefor and recording medium having program recorded thereon

The computational complexity of a denoising filter is reduced while suppressing a reduction in coding efficiency. In a video encoding/decoding apparatus which uses a loop filter, a deviation degree calculating unit calculates the degree of deviation between a denoising target pixel and a surrounding pixel using a decoded image. A template shape setting unit limits a template shape so that its size is proportional to the degree of deviation relative to the maximum value of the degree of deviation in the decoded image. When noise of the target pixel is removed, weighted in accordance with the degree of similarity between a template of the target pixel and a template of each of search points in a search shape and the weighted sum of pixel values at the search points, the loop filter performs template matching using the limited template shape and removes noise of the target pixel.

Video encoding/decoding methods, video encoding/decoding apparatuses, and programs therefor

In an image processing apparatus, an edge direction detection unit calculates an edge direction using a denoising target image. A search shape setting unit sets a search shape in which the number of search points along the direction perpendicular to the edge direction is smaller than the number of search points along the edge direction, using the calculated edge direction as an index used for limiting the search shape. When noise of a target pixel is removed with a weight in accordance with the degree of template similarity between a template for the target pixel and a template for each of search points within a search shape and the weighted sum of pixel values at the search points, a filter execution unit executes a process of performing template matching on only search points within the set search shape and removing the noise of the target pixel.

Image processing method, image processing apparatus, video encoding/decoding methods, video encoding/decoding apparatuses, and non-transitory computer-readable media therefor that perform denoising by means of template matching using search shape that is set in accordance with edge direction of image

An automatic producing method for a predicted value generation procedure that predicts a value of an encoding target pixel by using a previously-decoded pixel. A parent population is generated by randomly producing predicted value generation procedures each of which is indicated by a tree structure, and a plurality of predicted value generation procedures are selected as parents from the parent population. One or more predicted value generation procedures are generated as children based on a predetermined tree structure developing method which subjects the selected predicted value generation procedures to a development where an existing predicted value generation function can be an end node of a tree. A predicted value generation procedure having a best estimation cost is selected from among the predicted value generation procedures as the parents and children, where the sum of an information content for representing the tree structure and an amount of code evaluated by a predicted pixel value obtained by the tree structure is used as an estimation cost. A final predicted value generation procedure is produced by iterating the relevant operation.

Predicted pixel value generation procedure automatic producing method, image encoding method, image decoding method, apparatus therefor, programs therefor, and storage media which store the programs

Growing community of inventors

Yokosuka, Japan

Masaaki Matsumura