Electronics and Telecommunications Research Institute

Sanghun Yoon

Tae Moon Roh

Jung Hee Suk

Chun-Gi Lyuh

Ik Jae Chun

Sung-Jea Ko

You-sun Kim

Disclosed is a feature vector classification device which includes an initial condition setting unit; a variable calculating unit configured to receive a training vector and to calculate an error and a weight according to setting of the initial condition setting unit; a loop deciding unit configured to determine whether re-calculation is required, based on a comparison result between the calculated error and an error threshold; and a hyperplane generating unit configured to generate a hyperplane when an end signal is received from the loop deciding unit.

Feature vector classification device and method thereof

An image recognition device in accordance with the inventive concept may include an input vector extraction part extracting an input vector from an input image; a compression vector conversion part converting the input vector into a compression vector using a projection vector; a training parameter generation part receiving a training vector to generate a training parameter using a projection vector obtained through a folding operation of the training vector; and an image classification part classifying the compression vector using the training vector to output image recognition data.

Image recognition device and method of recognizing image thereof

Disclosed is an operating method of a direct memory access (DMA) controller having first and second DMA channels. The operating method includes iteratively performing a DMA transfer operation of the first DMA channel based upon loop information and transfer information of the first DMA channel; iteratively performing a DMA transfer operation of the second DMA channel based upon loop information and transfer information of the second DMA channel; reconfiguring the transfer and loop information of the first and second DMA channels; and again performing the iteratively performing a DMA transfer operation of the first DMA channel and the iteratively performing a DMA transfer operation of the first DMA channel based upon the reconfigured transfer and loop information of the first and second DMA channels.

Direct memory access controller and operating method thereof

Disclosed is an image registration device which includes an image input unit which receives an image; an image information generating unit which generates a homography matrix from the input image; and a warping unit which registers an image based on the homography matrix. The registration information generating unit comprises a distance information generator which generates distance information on subjects of the input image; a distance information modeler which approximates the generated distance information; an overlap information generator which generates overlap information from the approximated distance information; a matching pair determiner which determines a matching pair from the overlap information; and a homography matrix generator which generates a homography matrix from the matching pair.

Image registration device and method thereof

Provided is an image matching method of matching at least two images. The image matching method extracts feature points of a reference image and feature points of a target image, changes a feature point, selected from among the feature points of the reference image, to a reference point in the target image, sets a matching candidate region on the basis of the reference point, in the target image, and performs a similarity operation between the selected feature point in the reference image and a plurality of feature points included in the matching candidate region among the feature points of the target image. The image matching method decreases the number of similarity operations performed in the image matching operation, thereby guaranteeing a high-speed operation.

Image matching devices and image matching methods thereof

Provided are an object detection device and system. The object detection device includes an outline image extraction unit, a feature vector calculation unit, and an object judgment unit. The outline image extraction unit extracts an outline image from an input image. The feature vector calculation unit calculates a feature vector from the outline image by using histogram of oriented gradients (HOG) representing a frequency distribution of gradient vectors with respect to pixels of the outline image, and pixel coordinate information varying according to a spatial distribution of the gradient vectors. The object judgment unit judges a target object corresponding to the feature vector with reference to pre-learned data.

Growing community of inventors

Daejeon, South Korea

Sanghun Yoon