Fdn. for Res.&bus., Seoul Nat. Univ. of Sci.&tech.

Bong-Seok Seo

Dong Ho Kim

Changjong Hyun

Subin Hwang

Disclosed are an LSTM based personalization viewpoint estimation apparatus and method. According to an embodiment of the present disclosure, a viewpoint of a user is estimated using LSTM with respect to a 360-degree raw image and a viewpoint image that is a region of interest of the user and a tile including the estimated viewpoint is transmitted over a network, so that the amount of data to transmit compared to the network bandwidth may be reduced and a viewpoint of each individual may be estimated.

LSTM based personalization view point estimation apparatus and method

The present invention relates to a system and a method of preloading a multi-view video. According to an embodiment of the present invention, load of a system is reduced and client's Quality of Experience (QoE) is maximized. In addition, the sense of direction, distance, and space for virtual reality are felt in the same manner as those for the real environment, so that a virtual reality service is provided realistically, thereby further improving immersion and interest in the virtual reality service.

System and method for preloading multi-view video

An apparatus and method of producing sound using a terminal is proposed. According to an embodiment of the present invention, it is possible to provide realistic sound when playing contents and easily produce contents with improved immersion by making uniform a sound received through the microphone of the lightweight terminal through distortion compensation and recording the uniform sound in a stereophonic manner.

Apparatus and method for sound production using terminal

According to a specific embodiment of the present invention, an object at each viewpoint is recognized through image recognition, each recognized object is labeled, directional values and label values of objects that are determined as being the same are then input into an MPD, the MPD is transported, a direction of an object at a viewpoint to be moved, within an image, is corrected with a derived offset directional value, reproduction is performed, and thus, only with the labeling and the directional value of the object, the same content can be displayed on a screen without the user switching between directions. Accordingly, a reduction in an amount of data to be processed can improve a processing speed, and in turn, a sense of reality can be improved. Additionally, compatibility with an existing DASH structure is ensured, and applicability to existing image transport systems is widely expandable.

Growing community of inventors

Gyeonggi-do, South Korea

Bong-Seok Seo