An image may be evaluated by a computer vision system to determine whether it is fit for analysis. The computer vision system may generate an embedding of the image. An embedding quality score (EQS) of the image may be determined based on the image's embedding and a reference embedding associated with a cluster of reference noisy images. The quality of the image may be evaluated based on the EQS of the image to determine whether the quality meets filter criteria. The image may be further processed when the quality is sufficient, or otherwise the image may be removed.

A method for motion tracking and text recognition, the method including a step of generating ultrasound waves with a transmitter; a step of receiving the ultrasound waves at a receiver, the receiver including sensors that record the ultrasound waves; a step of estimating with a processor, angle-of-arrival information for the ultrasound waves; a step of associating the angle-of-arrival information with a gesture; a step of extracting features from the gesture; and a step of classifying the gesture as a specific text character based on the extracted features by comparing the extracted features with known text characters stored in one or more templates.

Systems and methods for performing spatial line grouping on digital ink stokes. The system includes an electronic processor configured to access a set of hypothetical lines in an electronic document and determine a set of hypothetical line pairings. The electronic processor is also configured to determine, via a gradient boosting tree model, a merge confidence score for each hypothetical line pairing and compare a first merge confidence score with a merge threshold. The first merge confidence score is associated with a first hypothetical line and a first neighboring hypothetical line. The electronic processor is also configured to, in response to the first merge confidence score satisfying the merge threshold, merge the first hypothetical line and the first neighboring hypothetical line to form a first line grouping. The electronic processor is also configured to perform a digital ink stroke analysis on the electronic document based on the first line grouping.

A coordinate input processing apparatus includes a position detection apparatus and a communication circuit. The position detection apparatus includes a sensor which detects a position pointed to by an electronic pen, and circuitry which acquires pen state information regarding a state of the electronic pen held by a person. The communication circuit transmits to an emotion estimation apparatus coordinates corresponding to the position pointed to by the electronic pen and the pen state information in an emotional state estimation request, and receives from the emotion estimation apparatus the coordinates corresponding to the position pointed to by the electronic pen, the pen state information included in the emotional state estimation request, and the information regarding the distracted state of the person holding the electronic pen in an emotional state estimation response having the same format as the emotional state estimation request.

Described techniques for multi-label classification, in which sequential data includes characters that have two or more aspects that require classification, are capable of providing separate classifications for different categories of components. Using an appropriately-trained neural network, the described techniques perform aligning and otherwise combining two or more classifications (e.g., categories, or types of labels) to obtain multi-label characters.

A computer system is provided that includes one or more processors configured to receive user input for inked content to a digital canvas, and process the inked content to determine one or more writing regions. Each writing region includes recognized text and one or more document layout features associated with that writing region. The one or more processors are further configured to tokenize a target writing region of the one or more writing regions into a sequence of tokens, process the sequence of tokens of the target writing region using task extraction subsystem that operates on tokens representing both the recognized text and the one or more document layout features of the target writing region, segment the target writing region into one or more sentence segments, and classify each of the one or more sentence segments as a task sentence or a non-task sentence.

Systems and methods for performing spatial line grouping on digital ink stokes. The system includes an electronic processor configured to access a set of hypothetical lines in an electronic document and determine a set of hypothetical line pairings. The electronic processor is also configured to determine, via a gradient boosting tree model, a merge confidence score for each hypothetical line pairing and compare a first merge confidence score with a merge threshold. The first merge confidence score is associated with a first hypothetical line and a first neighboring hypothetical line. The electronic processor is also configured to, in response to the first merge confidence score satisfying the merge threshold, merge the first hypothetical line and the first neighboring hypothetical line to form a first line grouping. The electronic processor is also configured to perform a digital ink stroke analysis on the electronic document based on the first line grouping.

Methods and systems for testing base text direction (BTD) include receiving one or more images captured by an end-user system. Each of the one or more images displays respective text test case information. Each of the one or more images is compared to a respective reference image associated with a respective text test case. It is determined whether the end-user system produces BTD errors based on the comparison in accordance with one or more BTD error rules.

The invention concerns a method comprising: detecting strokes of digital ink input on a computing device in a free handwriting format; detecting a text block from said strokes; performing text recognition on each text line of said text block, including extracting text lines from the text block and generating model data that associate each stroke of the text block with a character, a word and a text line of the text block; normalizing each text line from the free handwriting format into a structured format to comply with a document pattern. The normalization may comprise for each text line: computing a transform function to transform said text line into the structured format; applying the transform function to the text line; and updating the model data of said text line based on the transform function.

A training method, system, and computer program product include training a neural network including at least one of using norm-pooling as a non-linear function, using a two-sided ReLU as a non-linear function, and increasing a confidence gap and further training such that the network comprises a non-expansive network.