Image conversion

About: Image conversion is a research topic. Over the lifetime, 2490 publications have been published within this topic receiving 19077 citations.

Image processing technology and method based on deep learning

Abstract: The present invention relates to a computer structure combined with a neural network circuit and a parallel processing processor, and an image conversion and analysis processing method using the same. The computer structure of the present invention comprises: a neural network circuit for extracting information on a feature point, color, etc. for an image of an object, converting the same into a two-dimensional coordinate value for the image of the object, and outputting the coordinate value in a form of a packet; a plurality of processing units; a parallel processing processor for performing a multiple recognition circuit operation between pixels within an image patch formed around the extracted feature point; and an on-chip network apparatus connected between the neural network circuit and the parallel processing processor, and transmitting the packet for the coordinate values converted in the neural network circuit to the parallel processing processor. According to the present invention, a schematic process of the entire image is accelerated by a neural network circuit, so a detailed image processing is performed by a parallel processing digital processor of a high performance on the basis thereof. As a result, an existing image can be analyzed and converted into a replaced image.

2D to 3D Image Conversion Device and Method

Abstract: A method to convert two-dimensional (“2D”) image content into three-dimensional (“3D”) image content for display on a display device, comprises the steps of: analyzing the 2D image content for predefined indicators and generating a depth map for each of the predefined indicators; determining a combined depth map as a function of the generated depth maps; and generating the 3D image content for display on the display device as a function of the combined depth map.

Radiographic image pickup device

Abstract: PROBLEM TO BE SOLVED: To provide a highly precise enlarged radiographic image. SOLUTION: The radiographic image based on the radiation from a microfocus radiation source 11 transmitted through a subject 14 is recorded in a radiographic image conversion panel using an accelerated phosphorescent phosphor or a reader 30 using a flat panel detector. A space between the subject 15 and the reader 30 is variable, and the read gain of the radiographic image recorded in the reader 30 is set according to this space. A read picture element size according to the size of half-shadow of the focus diameter of a radiation source 11 is set. The image data obtained by reading of image is subjected to image processing such as processing for generating image data of enlarge image, processing for preventing the influence caused by half-shadow, profile emphasizing process in phase difference radiographing, or position correcting processing for correcting the boundary position detected in the phase difference radiographing to a right position.

Operation supporting system, image processing device, and deviation detecting method

Abstract: PROBLEM TO BE SOLVED: To provide an operation supporting system, an image processing device, and a deviation detecting method which more accurately determine whether deviation occurs in a position or a direction of a vehicle mounted camera or not in the operation supporting system which composes projected images which are projected on a plane of photographed images using a plurality of vehicle mounted cameras SOLUTION: The operation supporting system of the invention is provided with a first vehicle mounted camera which acquires a first image, a second vehicle mounted camera which acquires a second image, and an image conversion unit which converts the first image and the second image to a first projection image and a second projection image projected on the plane It is also provided with a solid object determination unit which determines whether there is a solid object in an overlapping region or not in a state that directions of the first vehicle mounted camera and the second vehicle mounted camera are different, and they have the overlapping region which overlaps mutually; a difference calculating unit which calculates the difference between an image in the overlapping region in the first projection image and an image in the overlapping region in the second projection image in a case that there is no solid state object; and a detection unit which detects that deviation occurs in the vehicle mounted cameras in a case that the difference is more than a threshold value COPYRIGHT: (C)2008,JPO&INPIT