The present invention provides a similarity search engine for a digital visual object (i.e., a digital image that represents a design, graphics, logo, symbols, words, or any combination thereof). The similarity search engine is based on a method that consists of conducting several independent search queries, thus each query examining a different aspect of similarity.

Computer-implemented systems and methods enhance a user’s sophistication as she/he reviews complex information sources using specialized detective tools provided by a user interface of the computer system. The specialized investigative inquiries are stored in a database and are particularly tailored a priori by a subject-matter content designer for the type of documents being reviewed for risk and opportunity. The investigative scripts are organized into to a path of risk-related subjects or topics, and within each path of subjects/topics the investigative scripts are organized into a specialized inquiry or flow chart.

An image processing apparatus includes an input unit configured to input image data, a learning unit configured to perform machine learning processing using information contained in the image data input by the input unit, an estimation unit configured to output an estimation result based on the information contained in the image data using a learning model generated by learning of the learning unit, and a determination unit configured to determine whether the image data input by the input unit contains sensitive information, wherein in a case where the determination unit determines that the image data input by the input unit contains the sensitive information, the learning unit does not perform machine learning on at least the sensitive information contained in the image data.

Systems and methods for generation and use of document analysis architectures are disclosed. A model builder component may be utilized to receiving user input data for labeling a set of documents as in class or out of class. That user input data may be utilized to train one or more classification models, which may then be utilized to predict classification of other documents. Trained models may be incorporated into a model taxonomy for searching and use by other users for document analysis purposes.

A number plate information specifying device includes an image acquisition unit that acquires a number plate image, a feature point extraction unit that extracts a feature point from the number plate image, a degree of similarity calculation unit that references a learning data set in which a plurality of feature points are recorded in association with a plurality of pieces of number plate information and calculates degrees of similarity for the feature points recorded in the learning data set that correspond to the feature point extracted from the number plate image, a vote value calculation unit that, on the basis of the degrees of similarity, calculates vote values for the pieces of number plate information recorded in the learning data set, and a specifying unit that specifies the piece of number plate information that has the highest vote value as the number plate information displayed in the number plate image.

Today, artificial neural networks are trained on large sets of manually tagged images. Generally, for better training, the training data should be as large as possible. Unfortunately, manually tagging images is time consuming and susceptible to error, making it difficult to produce the large sets of tagged data used to train artificial neural networks. To address this problem, the inventors have developed a smart tagging utility that uses a feature extraction unit and a fast-learning classifier to learn tags and tag images automatically, reducing the time to tag large sets of data. The feature extraction unit and fast-learning classifiers can be implemented as artificial neural networks that associate a label with features extracted from an image and tag similar features from the image or other images with the same label. Moreover, the smart tagging system can learn from user adjustment to its proposed tagging. This reduces tagging time and errors.

A device includes an image sensor and a processor to: receive training images that each include multiple visual features and an ROI; receive indications of locations of ROIs within each training image; perform at least one transform on the multiple visual features and ROI of at least one training image to align the multiple visual features and ROIs among all of the training images to identify a common set of visual features present within all of the training images; derive a converged ROI from at least a portion of the ROI of at least one training image; and generate an anchor model based on the converged ROI and the common set of visual features, wherein the common set of visual features defines the anchor and are each specified relative to the converged ROI, and the anchor model is used to derive a location of a candidate ROI in an image.

A graphic recognition-based interactive method and electronic device. The interactive method comprises: acquiring a first image to be analyzed (101); recognizing the first image to be analyzed to obtain a recognition result (102); playing a narration audio file associated with the recognition result (103); playing a question audio file associated with the interactive content of the first image to be analyzed based on the recognition result after the narration audio file is played (104); acquiring a second image to be analyzed (105); determining whether a feature graphic exists in the second image to be analyzed (106); if the feature graphic exists in the second image to be analyzed, determining whether the feature graphic overlaps with an interactive region, wherein the interactive region is an area corresponding to an answer to the question in the question audio file (108); playing a correct audio file if the feature graphic and the interactive region are overlapped (109); and playing a wrong audio file if the feature graphic and the interactive region are not overlapped (110). The said method provides increased interactivity.

A device includes an image sensor and a processor to: receive training images that each include multiple visual features and an ROI; receive indications of locations of ROIs within each training image; perform at least one transform on the multiple visual features and ROI of at least one training image to align the multiple visual features and ROIs among all of the training images to identify a common set of visual features present within all of the training images; derive a converged ROI from at least a portion of the ROI of at least one training image; and generate an anchor model based on the converged ROI and the common set of visual features, wherein the common set of visual features defines the anchor and are each specified relative to the converged ROI, and the anchor model is used to derive a location of a candidate ROI in an image.

A device may receive image data that includes an image of a document and lexicon data identifying a lexicon, and may perform an extraction technique on the image data to identify at least one field in the document. The device may utilize form segmentation to automatically generate label data identifying labels for the image data, and may process the image data, the label data, and data identifying the at least one field, with a first model, to identify visual features. The device may process the image data and the visual features, with a second model, to identify sequences of characters, and may process the image data and the sequences of characters, with a third model, to identify strings of characters. The device may compare the lexicon data and the strings of characters to generate verified strings of characters that may be utilized to generate a digitized document.