Abstract: Disclosed is a method of predicting crime occurrence in a prediction target region using big data, the method including: collecting crime prediction data of the prediction target region from a plurality of data domains; collecting crime occurrence data of crime that occurred in the prediction target region during a preset period of time from a crime occurrence record domain; analyzing the crime prediction data and the crime occurrence data of each of the data domains according to a statistical analysis, and extracting meaningful data of the crime prediction data as available data by each of the data domains; and predicting crime occurrence by applying the available data to a pre-registered deep learning algorithm, wherein the available data is classified into a plurality of data groups according to a data type, and the deep learning algorithm includes: a first deep neural network; a second deep neural network; and an output.

Abstract: There is provided an apparatus to measure stereovision including a frame body, a 3D display unit, a reflective mirror, an eyepiece, a half mirror, a stage, a plurality of protrusion degree markers, a target, and a target driving unit. A reflection angle of the half mirror is adjusted such that an imaginary image of an object having a protrusion degree of 0 in the 3D image reflected on the half mirror is located above the protrusion degree marker having a protrusion degree of 0, the imaginary image being seen through the eyepiece. The target driving unit is controlled such that the target corresponds to an imaginary image of an object of a 3D image having a particular protrusion degree. The stereovision of a tester is evaluated based on a protrusion degree marker to which the target is adjacent among the protrusion degree markers.

Abstract: The method for extracting salient object from stereoscopic video includes: dividing regions based on the similarity of color and the distance between pixels in a left-eye image and a right-eye image which are used for an input stereoscopic image; creating a disparity map based on a disparity obtained from a pixel difference of the left-eye image and the right-eye image; calculating a contrast-based saliency by comparing the divided regions and the divided regions of the disparity map; calculating a prior-knowledge-based saliency based on a prior-knowledge for the divided regions and the divided regions of the disparity map; and extracting salient regions of the image based on the contrast-based saliency and the prior-knowledge-based saliency.

Abstract: The invention relates to an apparatus for measuring stereovision comprising: a frame body having a space therein; a 3D display unit which is arranged on the upper part of the frame body and which display a 3D image downwards through the space; a reflective mirror which is arranged on the lower part of the frame body and which reflects the 3D image displayed from the 3D display unit in a forward direction; an eyepiece which is arranged between the 3D display unit and the reflective mirror on the rear part of the frame body and through which a 3D image is seen; a half mirror by which the 3D image reflected forward by the reflective mirror is penetrated and reflected and which is rotatably installed on the frame body to adjust a reflection angle upward or downward; a stage which is arranged in front of the half mirror, a plurality of protrusion degree markers having a predetermined protrusion degree being arranged on the bottom surface successively with a predetermined distance, a target being installed on the s

Abstract: The method for extracting salient object from stereoscopic video includes: dividing regions based on the similarity of color and the distance between pixels in a left-eye image and a right-eye image which are used for an input stereoscopic image; creating a disparity map based on a disparity obtained from a pixel difference of the left-eye image and the right-eye image; calculating a contrast-based saliency by comparing the divided regions and the divided regions of the disparity map; calculating a prior-knowledge-based saliency based on a prior-knowledge for the divided regions and the divided regions of the disparity map; and extracting salient regions of the image based on the contrast-based saliency and the prior-knowledge-based saliency.

Abstract: The method of face model alignment on unseen face includes inputting a face image; warping the face image into a standard shaped face based on a model seen by AAM (Active Appearance Model); normalizing the warped image by removing a texture change of the warped image; extracting a face texture from the normalized image; calculating error areas by comparing the face texture with the seen model; and aligning edges of the face texture with edges of the seen model while reducing the difference of the error areas.

Abstract: The invention relates to a method for measuring recognition warping with respect to a three-dimensional (3D) image. The method according to the invention comprises: (a) a step of selecting any one from a plurality of 3D parameters for a scale for recognizing a 3D image in a 3D manner; (b) a step of displaying a test stereoscopic image corresponding to the selected 3D parameter on a screen; (c) a step of displaying, on a screen, a plurality of sample images obtained by varying the value of the selected 3D parameter for the test stereoscopic image; (d) a step of selecting at least one of the plurality of sample images; and (e) a step of performing the steps (b) to (d) at least two times on different test stereoscopic images, and measuring a 3D recognition scale for the selected 3D parameter based on the sample image selected in the step (d).

Abstract: The invention relates to a stereovision optometry apparatus comprising: a 3D display unit for displaying 3D image; a reference 2D display unit for displaying reference 2D image; an eyepiece in opposite side of the 2D display unit configured to view the images; a reflective mirror in opposite side of the 3D display unit configured to move toward and away from the 3D display unit; and a half mirror arranged between the 3D display and the reflective mirror and disposed at a point where propagation paths of the 2D image and the 3D image intersect; wherein the 2D image penetrates through the half mirror to reach the eyepiece; wherein the 3D image penetrates the half mirror to be reflected on the reflective mirror and then the reflected image is reflected on the half mirror to reach the eyepiece.

Abstract: A method is provided that controls the color of a feeling-based image reflecting a user's preference via an interactive genetic algorithm. The method is performed in such a manner that a range of candidate colors is defined by type of feelings as templates, one from among the candidate templates, as the final template, is selected by learning a user's color preference via an interactive genetic algorithm, and then the color of an image is controlled based on the final template by recognizing a user's feeling.

Abstract: A visual communication method adapted to microblogs is provided. The visual communication method comprises analyzing, by a server, a user's input text using context and words, extracting, by the server, a picture corresponding to the text using a multimedia classification system data base, creating, by the server, an avatar representing the user, synthesizing, by the server, the picture and the avatar into one multimedia form, and transferring, by the server, the synthesized multimedia to the user's microblog.