scispace - formally typeset
Search or ask a question
Topic

Image conversion

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


Papers
More filters
Patent
21 Nov 1995
TL;DR: In this article, the authors proposed a device positioning system consisting of a first camera 3 for picking up the image of a circuit board 9 mounted on an X-Y table 2, a second camera 4 to pick up the images of a device 10 to be mounted on the circuit board and a monitor 6 for displaying an image based on an image signal delivered form the image processing section 5.
Abstract: PROBLEM TO BE SOLVED: To provide a device positioning system in which a device can be set conveniently at an accurate mounting position on a circuit board in a short time. SOLUTION: The device positioning system comprises a first camera 3 for picking up the image of a circuit board 9 mounted on an X-Y table 2, a second camera 4 for picking up the image of a device 10 to be mounted on the circuit board 9, an image processing section 5, and a monitor 6 for displaying an image based on an image signal delivered form the image processing section 5. The image processing section 5 produces an image signal indicative of a land formed on the circuit board 9 through image conversion of a board image signal delivered from the first camera 3 and an image signal indicative of the outline of device 10 through image conversion of a device image signal delivered from the second camera 4. Furthermore, the image processing section 5 synthesizes the land image signal and device outline image signal and delivers a synthesized image signal to the monitor 6 thus displaying a synthesized image. COPYRIGHT: (C)1997,JPO

9 citations

Patent
15 Jun 1998
TL;DR: In this article, an image analysis part judges how far image data of a negative image are from a gray line, and finds an intensity value of one optional color which provides a point on the gray line when combined with each of other colors, to identify an HP (highlight point) and an SP (shadow point) on the negative image.
Abstract: A positive image which resembles an object is obtained from a negative image which has various color distributions. An image analysis part judges how far image data of a negative image are from a gray line, and finds an intensity value of one optional color which provides a point on the gray line when combined with each of other colors, to thereby identify an HP (highlight point) and an SP (shadow point) on the negative image. At this stage, a pixel at which the sum of RGB-values on the negative image is approximately the largest is defined as a first HP candidate. Maximum values in the negative image are calculated for planes of the respective colors, and one set of the RGB-values is defined as a second HP candidate. Whether the first HP candidate is close to the gray line is judged. If the first HP candidate is close to the gray line, the first HP candidate is directly used as an HP. If the first HP candidate is far from the gray line, whether the second HP candidate is close to the gray line is judged. If the second HP candidate is close to the gray line, the second HP candidate is directly used as an HP. If the second HP candidate is far from the gray line, RGB-values in a white and a black regions in a gray scale are used as an HP and an SP. An image conversion part converts the negative image which is read from an image holding part into image data of a positive image which has a proper contrast, based on information regarding the HP and the SP on the negative image supplied from the image analysis part.

9 citations

Proceedings ArticleDOI
TL;DR: It was found that, for test target data of modest contrast, the resulting SFR measurements were only moderately sensitive to the use of the inverse OECF transformation.
Abstract: Measurement of the spatial frequency response (SFR) of digital still cameras by slanted-edge analysis has been established for several years. The method, described in standard ISO 12233, has also been applied to other image acquisition subsystems such as document and print scanners. With the frequent application of the method and use of supporting software, questions often arise about the form of the input test image data. The tone-transfer characteristics of the system under test can influence the results, as can signal quantization and clipping. For this reason, the original standard called for a transformation of the input data prior to the slanted-edge analysis. The transformation is based on the measured opto-electronic conversion function (OECF) and can convert the image data to a reference-exposure signal space. This is often helpful when comparing different devices, if the intent is to do so in terms of the performance of optics, detector, and primary signal processing. We describe the use of the OECF and its inverse to derive the signal transformation in question. The influence of typical characteristics will be shown in several examples. It was found that, for test target data of modest contrast, the resulting SFR measurements were only moderately sensitive to the use of the inverse OECF transformation.

9 citations

Patent
27 Jul 2018
TL;DR: In this paper, a uniform generative adversarial network-based multi-domain image conversion technology is proposed, which comprises the following processes of: learning to discriminate true and false images by using a discriminator D and classifying true images into a corresponding domain; taking the images and a target domain label as inputs of a generator G to generatea false image; under the condition of giving an original domain label, trying to reconstruct an original image by G according to the false image, and finally trying to generate an image which is not different from the true image and can be classified
Abstract: The invention discloses a uniform generative adversarial network-based multi-domain image conversion technology. The technology comprises the main contents of: training a discriminator; converting from an original domain to a target domain; converting from the target domain to the original domain; and sheltering the discriminator. The technology comprises the following processes of: learning to discriminate true and false images by using a discriminator D and classifying true images into a corresponding domain; taking the images and a target domain label as inputs of a generator G to generatea false image; under the condition of giving an original domain label, trying to reconstruct an original image by G according to the false image; carrying out continuous learning by D to discriminatetrue images and synthesized images, and carrying out continuous learning by G to shelter D; and finally trying to generate an image which is not different from the true image and can be classified into the target domain by D by G. According to the technology, a model is used for executing conversion from multi-domain images to images, so that the image conversion quality is improved, and the ability of flexibly converting input images into expected target images is provided.

9 citations

Patent
18 Jan 1999
TL;DR: In this paper, a multi-ap configuration filtering process without requiring the higher velocity operation of a memory in use is presented, where sampled input image signals are respectively coupled for each of a plurality of continuous pixel data and outputted by a coupled circuit.
Abstract: An image conversion process by a multitap configuration filtering process without requiring the higher velocity operation of a memory in use, wherein sampled input image signals are respectively coupled for each of a plurality of continuous pixel data and outputted by a coupled circuit (102). Approximately one frame of the output data of the coupled circuit (102) are stored in a memory device (103). The memory device (103) outputs signals to a memory device (104) in accordance with the output of a memory device (107) in which control data calculated beforehand are stored. The memory device (104) stores signals inputted from the memory device (103) larger in number than taps of a filtering circuit (106) and outputs them to a selection circuit (105). The selection circuit (105) selects signals required by the filtering circuit (106) from the signals inputted from the memory device (104) and outputs them to the filtering circuit (106). The filtering circuit (106) generates output image signals by using the signals inputted from the selection circuit (105).

9 citations


Network Information
Related Topics (5)
Image processing
229.9K papers, 3.5M citations
84% related
Feature (computer vision)
128.2K papers, 1.7M citations
83% related
Pixel
136.5K papers, 1.5M citations
82% related
Feature extraction
111.8K papers, 2.1M citations
80% related
Image segmentation
79.6K papers, 1.8M citations
79% related
Performance
Metrics
No. of papers in the topic in previous years
YearPapers
202132
202074
2019117
2018115
2017100
2016107