Abstract: An image retrieval method applies an application server, one or more calculating servers, and a sorting server to perform image retrieval. The application server extracts visual features of an exemplary image. The one or more calculating servers calculate similarities of available images according to the visual features of the exemplary image. The sorting server sorts the available images according to the similarities so as to obtain images similar to the exemplary image.

Abstract: A method for isolating and cutting out figures in a design patent document divides each page of a figure section of the design patent document into a multiplicity of areas according to histograms, which are created based on information of black pixels and white pixels in a black-and-white image of the page. The method selects the areas that include the figures of the design patent, and displays the selected areas on a display device.

Abstract: A method for automatically typesetting patent images extracts a brief introduction of each patent image from a description part of a patent document, and records a keyword of the brief introduction. The method distinguishes an image label of each patent image from an image part of the patent document. The method rotates the patent image by ninety degrees clockwise in response that the image label of the patent image does not contain the keyword, and outputs the rotated image onto a display device.

Abstract: An image processing system and method for comparing and correcting two monochromic images A2 and B2 that extracts a skeleton of objects in the monochromic image A2 to generate a skeleton image A3 and extracts a skeleton of objects in the monochromic image B2 to generate a skeleton image B3. The system and method then covers the skeleton image A3 with the monochromic image B2 to generate a covered image A4, and covers the skeleton image B3 with the monochromic image A2 to generate a covered image B4. The system and method further corrects allowable variances in the covered images A4 and B4, and outputs the covered images whose variances have been corrected.

Abstract: An electronic document comparison system and method removes cachets and noise from a test electronic document. The system and method further compares each of second minimum blocks with a corresponding first minimum block in a standard electronic document line by line, and obtains the second minimum blocks having different coordinates on each line. Furthermore, the system and method simplifies the obtained second minimum blocks having different coordinates by filtering designated objects, and marks the simplified second minimum blocks in the test electronic document.

Abstract: An image processing system and method for comparing two monochromic images A2 and B2 is provided. The system and method thins objects in the monochromic image B2 so as to generate a skeleton image B3, thickens the objects in the monochromic image A2 to generate a bold image A4, and overlays the skeleton image B3 with the bold image A4 so as to generate an overlaid image AB1. The system and method further thins the objects in the monochromic image A2 so as to generate a skeleton image A3, thickens the objects in the monochromic image B2 to generate a bold image B4, and overlays the skeleton image A3 with the bold image B4 so as to generate an overlaid image AB2. The system and method outputs the overlaid images AB1 and AB2 on a display screen.

Abstract: An image retrieval method applies an application server, one or more calculating servers, and a sorting server to perform image retrieval. The application server extracts visual features of an exemplary image. The one or more calculating servers calculate similarities of available images according to the visual features of the exemplary image. The sorting server sorts the available images according to the similarities so as to obtain images similar to the exemplary image.

Abstract: An electronic device and method of displaying design patents receives search conditions, and searches at least one image of a design patent from a database server. An image is selected from the searched images, to acquire a design patent that contains the selected image and acquire all images of the design patent. A cubic box is created. The images of the design patent are superimposed onto the surfaces of the cubic box, to generate a cubic view of the design patent. Each of the search conditions and one or more cubic views corresponding to the search condition is recording into a search list. All contents of each of the design patents corresponding to the cubic views in a single page.

Abstract: A processing digital noise method generates a matrix according to pixel values of pixels of a digital image, and selects a sub-matrix from the matrix. The sub-matrix has the same dimensions as a digital noise sample and includes a pixel value of a pixel selected from the digital image. The digital noise sample is represented as a matrix composed of 0 and 1. The method compares the sub-matrix with the digital noise sample to determine whether or not all values in the sub-matrix and the digital noise sample are the same. If all values are the same, the method further searches 1 from the sub-matrix for locating digital noise of the digital image, and removes the digital noise of the digital image by changing the pixel values of the digital noise from 1 to 0.

Abstract: An image retrieval method applies an application server, one or more calculating servers, and a sorting server to perform image retrieval. The application server extracts visual features of an exemplary image. The one or more calculating servers calculate similarities of available images according to the visual features of the exemplary image. The sorting server sorts the available images according to the similarities so as to obtain images similar to the exemplary image.

Abstract: An image comparison method compares a reference image with a test image. Each image includes objects and a background. The method generates a skeleton image of the reference image. The skeleton image of the reference image is compared with the test image so as to determine if the reference image have more objects or objects parts than the test image. Similarly, a skeleton image of the test image is generated. The skeleton image of the test image is compared with the reference image so as to determine if the test image have more objects or objects parts than the reference image.

Abstract: A method for isolating and cutting out figures in a design patent document divides each page of a figure section of the design patent document into a multiplicity of areas according to histograms, which are created based on information of black pixels and white pixels in a black-and-white image of the page. The method selects the areas that include the figures of the design patent, and displays the selected areas on a display device.

Abstract: A method for cutting out a summary diagram of a patent document reads a first page of a patent document and divides the first page into multiple blocks. The method selects the block which has a width value greater than a predetermined width value, and cut off blank areas of the selected block, to maintain a area that includes the summary diagram in the selected block. The method displays the area as the diagram in a search result of the patent document on a display device, and the area contains all the text of the first page if no summary diagram is in the first page.

Abstract: A method for identifying components in figures extracts component numbers in a description section of a patent document that are coupled with a component name, and creates a component information list based on the component numbers. The method distinguishes each component number and respective positions of all the component numbers in a figure section of the patent document. The method detects a position of a cursor displayed on a display device when the figure section is being viewed. The method searches for a component name of a component number from the component information list when the cursor is positioned within a preset region around the component number, and displays the component name beside the component number in the figure section of the patent document on the display device.

Abstract: A method for automatically typesetting patent images extracts a brief introduction of each patent image from a description part of a patent document, and records a keyword of the brief introduction. The method distinguishes an image label of each patent image from an image part of the patent document. The method rotates the patent image by ninety degrees clockwise in response that the image label of the patent image does not contain the keyword, and outputs the rotated image onto a display device.

Abstract: A method for extracting component names from a document reads text content of the document, searches for component labels in the text content, and stores a position of each component label in the text content in a storage device. The method further extract a component name corresponding to each component label in the text content according to the position of each component label, and creates a component table according to the component label and the component name.

Abstract: An electronic document comparison system and method removes cachets and noise from a test electronic document. The system and method further compares each of second minimum blocks with a corresponding first minimum block in a standard electronic document line by line, and obtains the second minimum blocks having different coordinates on each line. Furthermore, the system and method simplifies the obtained second minimum blocks having different coordinates by filtering designated objects, and marks the simplified second minimum blocks in the test electronic document.

Abstract: A processing digital noise method generates a matrix according to pixel values of pixels of a digital image, and selects a sub-matrix from the matrix. The sub-matrix has the same dimensions as a digital noise sample and includes a pixel value of a pixel selected from the digital image. The digital noise sample is represented as a matrix composed of 0 and 1. The method compares the sub-matrix with the digital noise sample to determine whether or not all values in the sub-matrix and the digital noise sample are the same. If all values are the same, the method further searches 1 from the sub-matrix for locating digital noise of the digital image, and removes the digital noise of the digital image by changing the pixel values of the digital noise from 1 to 0.

Abstract: An image processing system and method for comparing two monochromic images A2 and B2 is provided. The system and method thins objects in the monochromic image B2 so as to generate a skeleton image B3, thickens the objects in the monochromic image A2 to generate a bold image A4, and overlays the skeleton image B3 with the bold image A4 so as to generate an overlaid image AB1. The system and method further thins the objects in the monochromic image A2 so as to generate a skeleton image A3, thickens the objects in the monochromic image B2 to generate a bold image B4, and overlays the skeleton image A3 with the bold image B4 so as to generate an overlaid image AB2. The system and method outputs the overlaid images AB1 and AB2 on a display screen.

Abstract: An image processing system and method for comparing and correcting two monochromic images A2 and B2 is provided. The system and method extracts a skeleton of objects in the monochromic image A2 to generate a skeleton image A3 and extracts a skeleton of objects in the monochromic image B2 to generate a skeleton image B3. The system and method then covers the skeleton image A3 with the monochromic image B2 to generate a covered image A4, and covers the skeleton image B3 with the monochromic image A2 to generate a covered image B4. The system and method further corrects allowable variances in the covered images A4 and B4, and outputs the covered images whose variances have been corrected.