Sony Corporation

Sony Group Corporation

Saturn Licensing LLC

Tamaki Kojima

Tomohiko Gotoh

Katsuhiro Takematsu

Yoshihiro Yamaguchi

Mikio Sakemoto

Makoto Murata

Takeshi Fukuda

Masatomo Kurata

Chao Xu

Daisuke Mochizuki

Yun Sun

Noriaki Nakagawa

Yuka Eura

Tsunayuki Ohwa

Kenichi Agata

Osamu Hanagata

Brian Clarkson

Nozomi Hara

Wenwu Zhao

Yoshiyuki Kobayashi

Yusuke Sakai

Yasuhisa Nakajima

Haruo Oba

Takuro Noda

Satoru Inoue

Mohammad Gharavi-Alkhansari

Masataka Suzuki

Kazuyoshi Suzuki

Takuo Ikeda

Naoki Okamoto

Masahiro Nakano

Hidenori Karasawa

Takanori Nishimura

Takashi Tsurumoto

Osamu Date

QiHong Wang

Yuuji Takimoto

Kazumoto Kondo

Hidenori Yamaji

Ryo Mukaiyama

Hiroyuki Takeda

Ryu Aoyama

Ichiro Ueno

Baiping Liao

Akira Tokuse

Akiko Terayama

Soichi Sato

Katsuhiro Hashimoto

Takaaki Morimura

Shojiro Sato

Shinichi Numata

YuHang Shen

Atsushi Miyairi

Yasuki Ogiwara

Suzuki Kazuyoshi

Provided is a system that receives a three-dimensional (3D) skinned model of a first object and a set of images of a second object. The second object is same as or structurally similar to the first object in appearance. The system executes a differentiable rendering operation to generate a first image which includes a contour of the 3D skinned model. The system further generates a reference image which includes a contour of the second object, based on the received set of images. The system updates first values of bone parameters associated with the 3D skinned model based on a difference between the generated first image and the generated reference image. The system updates the 3D skinned model based on the updated first values of the bone parameters.

Motion retargeting based on differentiable rendering

Recent progress of neural networks has shown significant progress over human pose estimation tasks. Pose estimation could be categorized into monocular 2D pose estimation, multi-view 3D pose estimation, and single view 3D pose estimation, where recently 3D pose is getting more attention to be applied to AR/VR, game and human computer interaction applications. However, current academic benchmarks on human 3D pose estimation only respect performance on their relative pose. The root positioning over the time, in another word, the 'trajectory' of the whole body in 3D space is not considered well enough. Applications such as motion capture, not only require precise relative pose of the body but also the root position of the whole body in 3D space. Therefore, an efficient monocular full 3D pose recovery model from 2D pose input is described herein, which is able to be applied to the above applications.

Full skeletal 3D pose recovery from monocular camera

An image processing system may include an imaging device for capturing an image and an image processing apparatus for processing the image. The imaging device may include an imaging unit for capturing the image, a first recording unit for recording information relating to the image, the information being associated with the image, and a first transmission control unit for controlling transmission of the image to the image processing apparatus. The image processing apparatus may include a reception control unit for controlling reception of the image transmitted from the imaging device, a feature extracting unit for extracting a feature of the received image, a second recording unit for recording the feature, extracted from the image, the feature being associated with the image, and a second transmission control unit for controlling transmission of the feature to the imaging device.

System, apparatus, method, program and recording medium for processing image

A system and method for semantic segmentation with a soft cross-entropy loss is provided. The system inputs a first color image to an input layer of a semantic segmentation network for a multi-class classification task. The semantic segmentation network generates, at an auxiliary stride, a first feature map as an output of an auxiliary layer of the semantic segmentation network based on the input first color image. The system extracts the generated first feature map from the auxiliary layer and computes a probability map as a set of soft labels over a set of classes of the multi-class classification task, based on the extracted first feature map. The system further computes an auxiliary cross-entropy loss between the computed probability map and a ground truth probability map for the auxiliary stride and trains the semantic segmentation network for the multi-class classification task based on the computed auxiliary cross-entropy loss.

Semantic segmentation with soft cross-entropy loss

An electronic device and method for image mask generation using a deep neural network is provided. The electronic device stores an input image that includes an object of interest generates a pilot mask for the input image by application of a trained deep neural network (DNN) on the input image. The electronic device extracts, from the input image, a region of interest that includes the object of interest, based on the generated pilot mask and resizes the extracted region of interest such that a size of the resized region of interest equals a threshold scaling size. The electronic device further generates a final mask by reapplication of the trained DNN on the resized region of interest.

Image mask generation using a deep neural network

A content management apparatus includes a content storage section configured to store a piece of content together with a content unique identifier; a content transmitter configured to transmit the piece of content together with the content unique identifier to a Web server; an intra-server management identifier receiver configured to receive an intra-server management identifier corresponding to the content unique identifier from the Web server; a management table for storing the content unique identifier and the intra-server management identifier; a content information updating message receiver configured to receive content information on the piece of content and a content information updating message containing the intra-server management identifier from the Web server; an identifier converter configured to convert the intra-server management identifier contained in the content information updating message into the content unique identifier; and a content management section configured to allow the content storage section to store the content information.

Content management apparatus, web server, network system, content management method, content information management method, and program

An information processing apparatus includes an extraction unit operable to extract metadata used for recognizing an event associated with a captured image from at least one of the captured image and attribute information concerning the captured image, and a recognition unit operable to recognize, on the basis of the metadata extracted by the extraction unit, the event associated with the captured image from among a plurality of predetermined events.

Information processing apparatus and method, and program

Provided is an image display device provided with an image display unit for displaying a still image or a moving image, an image capturing unit for capturing the image in a direction in which the image display unit displays the image on the image display unit, and a face detection unit for performing, at predetermined intervals, face detection processing for detecting a face included in the image captured by the image capturing unit, wherein the predetermined intervals are variable according to whether or not the face of a user is included in the image captured by the image capturing unit, and when the position of the detected face does not change for a predetermined period of time in the face detection processing, the face detection unit does not detect the face as the face of the user.

Image display device, control method and computer program

Growing community of inventors

Kanagawa, Japan

Tamaki Kojima