Abstract: Methods, apparatus, systems and articles of manufacture (e.g., physical storage media) to assess image quality using adaptive non-overlapping mean estimation are disclosed. Example apparatus disclosed herein include a machine learning system to be trained to classify image quality. Disclosed example apparatus also include a feature extractor to apply a blur filter to a first image to determine a blurred image, determine a blur feature value for the first image, the blur feature value to represent an amount the first image differs from the blurred image, and apply a vector of feature values associated with the first image to the machine learning system, the vector of feature values including the blur feature value. Disclosed example apparatus further include an image classifier to classify image quality associated with the first image based on an output of the machine learning system responsive to the vector of feature values associated with the first image.

Abstract: Image segmentation based on the combination of a deep learning network and a shape-guided deformable model is provided. In various embodiments, a time sequence of images is received. The sequence of images is provided to a convolutional network to obtain a sequence of preliminary segmentations. The sequence of preliminary segmentations labels a region of interest in each of the images of the sequence. A reference and auxiliary mask are generated from the sequence of preliminary segmentations. The reference mask corresponds to the region of interest. The auxiliary mask corresponds to areas outside the region of interest. A final segmentation corresponding to the region of interest is generated for each of the sequence of images by applying a deformable model to the composite mask with reference to the auxiliary mask.

Abstract: A system for real time bilinear interpolation includes a bilinear interpolation module capable of: generating pixel addresses for original image pixels of an original image needed for performing bilinear interpolation of the original image to form a resized image, wherein the generated pixel addresses assume all the original image pixels of the original image are accessible, and performing bilinear interpolation, and a pixel smart memory module capable: of sequentially receiving original image pixel rows of the original image an original image pixel row a time, predicting which original image pixel rows are needed for performing bilinear interpolation, storing only the needed sequentially received original image pixel rows in a memory, decoding the generated pixel addresses to form decoded addresses to access the needed original image pixel rows stored in the memory, and sending the needed original image pixel rows to the bilinear interpolation module for performing bilinear interpolation.

Abstract: A method for computer assisted identification of appropriate anatomical structure for placement of a medical device, comprising: receiving a 3D scan volume comprising set of medical scan images of a region of an anatomical structure where the medical device is to be placed; automatically processing the set of medical scan images to perform automatic segmentation of the anatomical structure; automatically determining a subsection of the 3D scan volume as a 3D ROI by combining the raw medical scan images and the obtained segmentation data; automatically processing the ROI to determine the preferred 3D position and orientation of the medical device to be placed with respect to the anatomical structure by identifying landmarks within the anatomical structure with a pre-trained prediction neural network; automatically determining the preferred 3D position and orientation of the medical device to be placed with respect to the 3D scan volume of the anatomical structure.

Abstract: Methods, systems, and apparatus for facial recognition. An example method includes storing encrypted facial features of a target user that are encrypted by using a first public key; receiving a recognition request from an end-user device of the target user, wherein the recognition request comprises an encrypted facial image and a second public key; performing homomorphic feature processing on the encrypted facial image; obtain obfuscated and encrypted output features; obtain obfuscated and encrypted facial features; transmitting the obfuscated and encrypted output features and the obfuscated and encrypted facial features to the end-user device; receiving from the end-user device a difference between the second intermediate value and the first intermediate value; and determining whether the to-be-recognized facial image corresponds to a facial image of the target user, comprising removing an impact of the first obfuscation and the second obfuscation on the difference.

Abstract: Disclosed herein are system, method, and computer program product embodiments for processing a document. In an embodiment, a document processing system may receive a document. The document processing system may perform optical character recognition to obtain character information and positioning information for the characters. The document processing system may generate a down-sampled two-dimensional character grid for the document. The document processing system may apply a convolutional neural network to the character grid to obtain semantic meaning for the document. The convolutional neural network may produce a segmentation mask and bounding boxes to correspond to the document.

Abstract: A client device and method to normalize a digital image into a universal digital color system includes a camera and a processor. The camera acquires a digital image of a color swatch and the processor converts the digital image into a RGB color image of the color swatch. The RGB color image is segmented into segments and each segment is analyzed to determine a dominant color for each segment. At least one dominant color for the color swatch is determined based on prevalence of at least one dominant color in each segment. A hexadecimal code of the universal color system is assigned to the color swatch that is closest to a digital hexadecimal value of the RGB image based on color component intensity values of the dominant color of the color swatch, thereby normalizing RGB colors of the color swatch.

Abstract: The Loupe system defines Loupe Visual Art DNA for art images to be presented to a user so as to maximize and customize the user experience in viewing art images delivered onto digital displays, TVs and other screens facilitating the artwork transition with and without human interaction. The Loupe system recommendations engine utilizes both human and machine curated data to determine factors of art images that will appeal to a user viewing the images. The Loupe system gathers data about visual perception, historical and academic provenance, and emotion or intention represented in an image. The gathered data is analyzed through deep learning and Al algorithms to inform recommendations and select art images to be presented to a user. The user may purchase fine art prints or select originals of the artwork image displayed, if the artist elects to make it available for sale, presented from the Loupe integrated electronic marketplace.

Abstract: Techniques are described for a de-obfuscation framework that utilizes image recognition of text. A word input by a user is received by the de-obfuscation service. Visual feature data associated with an image corresponding to each character of the word is generated. Word embeddings are generated using the visual feature data and each character of the word using a character encoder layer. Feature vectors are generated from the word embedding by combining the generated word embeddings and a provided word embedding using a second neural network. The generated feature vector is classified. Potential text obfuscation is detected from the classified generated feature vector using a lexicon to determine de-obfuscated text closet to the user text.

Abstract: This disclosure relates to systems and methods for solving generic inverse problems by providing a coupled representation architecture using transform learning. Convention solutions are complex, require long training and testing times, reconstruction quality also may not be suitable for all applications. Furthermore, they preclude application to real-time scenarios due to the mentioned inherent lacunae. The methods provided herein require involve very low computational complexity with a need for only three matrix-vector products, and requires very short training and testing times, which makes it applicable for real-time applications. Unlike the conventional learning architectures using inductive approaches, the CASC of the present disclosure can learn directly from the source domain and the number of features in a source domain may not be necessarily equal to the number of features in a target domain.

Abstract: Disclosed herein is a point cloud data reception device including a receiver configured to receive a bitstream containing geometry data of point cloud data and/or attribute data of the point cloud data, and/or a decoder configured to decode the point cloud data in the bitstream. Disclosed herein is a point cloud data transmission device including an encoder configured to encode point cloud data, and/or a transmitter configured to transmit a bitstream containing the point cloud data and/or signaling information about the point cloud data.

Abstract: A system for judging a situation of an object includes a component analysis unit that extracts a feature point in an image and analyzes at least two components selected from an object contained in the image, a posture, a shape, an orientation, and an expression of the object, a type and a location of a part of the object, and a background in the image; a situation acquisition unit that acquires situation data indicating the situation of the object from an image search computer; a learning unit that associates and learns a combination of the components with the acquired situation data; and an object situation judgement unit that judges the situation of the object based on the learning result of the situation data when a result of analysis for a predetermined image is same as or similar to the combination of the components.

Abstract: Systems and methods for selective use of examples are provided. For example, information associated with external devices may be obtained, weights may be assigned to examples using the obtained information, and inference models may be generated using the assignment of weights. In some examples, update information associated with the external devices may be obtained, the assignment of weights may be updated using the update information, and updated inference models may be generated using the updated assignment of weights.

Abstract: Systems and methods here may be used for pre-processing images, including using a computer for receiving a pixelated image of a paper document of an original size, downscaling the received pixelated image, employing a neural network algorithm to the downscaled image to identify four corners of the paper document in the received pixelated image, re-enlarging the downscaled image to the original size, identifying each of four corners of the paper document in the pixelated image, determining a quadrilateral composed of lines that intersect at four angles at the four corners of the paper document in the pixelated image, defining a projective plane of the pixelated image, and determining an inverse transformation of the pixelated image to transform the projective plane quadrilateral into a right angled rectangle.

Abstract: A method, computer program product, and system for generating and embedding a watermark in digital video frame include a processor obtaining a request to generate a watermark and embed the watermark in a digital video frame captured by a first monitor. Based on obtaining the request, the processor fetches from one or more pre-defined regions of a memory resource, digital video data captured by at least two monitors, where a timestamp of the digital video data is equal to a timestamp of the digital video frame. The processor generates a watermark from the digital video data by calculating a binary result of the digital video data. The processor embeds the watermark (binary result) in the digital video frame.

Abstract: A facial image processing method and apparatus, an electronic device and a computer readable storage medium. The facial image processing method comprises: acquiring an image to be processed; detecting whether a facial image exists in the image to be processed; in response to the detected facial image, carrying out a first processing on the facial image to obtain a first image; and obtaining depth information, and adding the depth information to the first image to obtain a processed image. According to the method, the image depth information can be added after the image is processed, such that the image has a stereoscopic impression and the user operation is simplified.

Abstract: A method of transmitting point cloud data in accordance with embodiments may include encoding point cloud data, and/or transmitting a bitstream containing the encoded point cloud data and metadata for the point cloud data. In addition, a method of receiving point cloud data in accordance with embodiments may include receiving a bitstream containing point cloud data and metadata, and/or decoding the point cloud data.

Abstract: An apparatus, method, system and computer-readable medium are provided for generating one or more descriptors that may potentially be associated with content, such as video or a segment of video. In some embodiments, a teaser for the content may be identified based on contextual similarity between words and/or phrases in the segment and one or more other segments, such as a previous segment. In some embodiments, an optical character recognition (OCR) technique may be applied to the content, such as banners or graphics associated with the content in order to generate or identify OCR'd text or characters. The text/characters may serve as a candidate descriptor(s). In some embodiments, one or more strings of characters or words may be compared with (pre-assigned) tags associated with the content, and if it is determined that the one or more strings or words match the tags within a threshold, the one or more strings or words may serve as a candidate descriptor(s).

Abstract: The orientation of imagery relative to a compass bearing may be determined based on the position of the sun or other information relating to celestial bodies captured in the image.

Abstract: Signal processing devices and methods obtain an estimate of a geometric transform of an image signal, use it to obtain a measure of signal confidence of a reference signal in a sub-block of the image signal, and weight message estimates based on the signal confidence. The geometric transform is used as an approximation of the geometric distortion of an image after digital data is encoded in it, and it is used to compensate for this distortion to facilitate extracting embedded digital messages from the image. Due to the errors in the approximation, the signal confidence metric is determined and used to weight message symbol estimates extracted from the image.