Abstract: A computer-implemented method for identifying utility usage from a historical utility file is disclosed. The method includes obtaining a file containing historical utility consumption of a dwelling over a time period; processing the file through optical character recognition (OCR); identifying contextual data from the OCR processed file; identifying chart data from the OCR processed file; extracting one or more values from the chart data, wherein the values correspond to one or more elements of the chart data; and contextualizing the extracted values from the chart data by applying the contextual data to the extracted value to obtain utility usage data.

Abstract: A method and a system for deriving visual rhythm from a video signal are described. A feature extraction module receives the video signal and extracts a two-dimensional feature from the video signal. A one-dimensional video feature computation module derives a one-dimensional feature from the extracted two-dimensional feature. A visual rhythm detector module detects a visual beat and a visual tempo from the one-dimensional feature.

Abstract: A method and a system for deriving visual rhythm from a video signal are described. A feature extraction module receives the video signal and extracts a two-dimensional feature from the video signal. A one-dimensional video feature computation module derives a one-dimensional feature from the extracted two-dimensional feature. A visual rhythm detector module detects a visual beat and a visual tempo from the one-dimensional feature.

Abstract: A semi-automatic system for scanning a document includes an image capture device, such as a digital or video camera, which records a sequence of images while a user waves a document in front of the device. The user can present multiple pages of a document to the image capture device, after which the total sequence of images are processed to identify a clear image of each page from the sequence of images. The system further includes image processing techniques to correct for motion blurring, acceleration and perspective errors. The system is capable of processing any size or shape of document without destroying the organization or format of the original.

Abstract: A method and a system for deriving visual rhythm from a video signal are described. A feature extraction module receives the video signal and extracts a two-dimensional feature from the video signal. A one-dimensional video feature computation module derives a one-dimensional feature from the extracted two-dimensional feature. A visual rhythm detector module detects a visual beat and a visual tempo from the one-dimensional feature.

Abstract: System and method for graphical user authentication using compact collages of regions of images. Image corpus is constructed by selecting similar quality images and filtering to further homogenize image quality. Regions are detected within the images and scored according to similarity and neighborhood information. Regions with lower scores provide less information about other regions and are more secure secrets. During enrollment, user selects secret images; decoy images are chosen by the system accordingly. Regions from secret images are selected as secrets according to the scoring of regions and regions from decoy images are selected as decoys. A collage is formed with secrets and decoys. Compact rendering enhances security and is suitable for small displays of mobile devices. Several rounds of challenge, requiring identification of secrets, are presented to the user. User is authenticated if a certain number of correct identifications of secrets within a number of rounds are achieved.

Abstract: A semi-automatic system for scanning a document includes an image capture device, such as a digital or video camera, which records a sequence of images while a user waves a document in front of the device. The user can present multiple pages of a document to the image capture device, after which the total sequence of images are processed to identify a clear image of each page from the sequence of images. The system further includes image processing techniques to correct for motion blurring, acceleration and perspective errors. The system is capable of processing any size or shape of document without destroying the organization or format of the original.

Abstract: System and method for graphical user authentication using compact collages of regions of images. Image corpus is constructed by selecting similar quality images and filtering to further homogenize image quality. Regions are detected within the images and scored according to similarity and neighborhood information. Regions with lower scores provide less information about other regions and are more secure secrets. During enrollment, user selects secret images; decoy images are chosen by the system accordingly. Regions from secret images are selected as secrets according to the scoring of regions and regions from decoy images are selected as decoys. A collage is formed with secrets and decoys. Compact rendering enhances security and is suitable for small displays of mobile devices. Several rounds of challenge, requiring identification of secrets, are presented to the user. User is authenticated if a certain number of correct identifications of secrets within a number of rounds are achieved.

Abstract: A watermarking method and device to embed a digitized watermark into a digital image file in the transformation domain is disclosed. In the watermarking technology of this invention, the digitized watermark is first error corrected to enhance its robustness. The digital image file to be embedded with the digitized watermark is first processed such that capacity of each element of the digital image file is calculated. The capacity of an element of the digital image file represents the highest value of tolerance for an element of the watermark to be added to the said element of the digital image file. Elements of the digital image file to which watermark elements may be added are selected and their addresses are recorded. The selected elements are then modulated so that they are suited to be embedded with elements of the digital watermark. After modulation the digitized watermark is embedded into the digital file by distributing the modulated elements of the watermark into the selected addresses.

Abstract: A watermarking method and device to embed a digitized watermark into a digital image file in the spatial domain first performs error correction to enhance the robustness of the digital image. The digital image file is first processed such that capacity of each element of the digital image file is calculated. The capacity represents the highest power of element of the watermark that can be added to the related element of the digital file. Elements to which watermark elements may be added are selected and their addresses are stored. Elements of the watermark file are then inserted into the selected addresses in sequence and the watermark is modulated. After the modulation the digitized watermark is embedded into the digital file by distributing the modulated elements of the watermark into the selected addresses.