The present disclosure is directed toward systems and methods that efficiently and effectively generate an enhanced document image of a displayed document in an image frame captured from a live image feed. For example, systems and methods described herein apply a document enhancement process to a displayed document in an image frame that result in an enhanced document image that is cropped, rectified, un-shadowed, and with dark text against a mostly white background. Additionally, systems and method described herein determine whether a stored digital content item includes a displayed document. In response to determining that a stored digital content item does include a displayed document, systems and methods described herein generate an enhanced document image of a displayed document included in the stored digital content item.

An apparatus, system and method for providing a visual search using augmented reality glasses. The apparatus, system and method include a network communicatively associated with the glasses capable of providing remote connectivity to an application programming interface (API); a machine learning (ML) model communicative with the API and having an input capable of receiving live video data indicative of a view field of the glasses, wherein the ML model includes at least a data comparator and platform-specific coding corresponded to the glasses; a search engine within the ML model and having a secondary input interfaced to a comparative database, wherein the search engine compares the live view field video data to the secondary input using the comparator; and a match output capable of outputting a match obtained by the search engine over the network to the glasses.

An artificial intelligence-based image processing method implemented by a computer device is provided. The method includes: acquiring an image; performing element region detection on the image to determine an element region in the image; detecting a target element region in the image using an artificial intelligence-based technique; generating a target element envelope region by searching an envelope for the detected target element region; and fusing the element region and the target element envelope region to obtain a target element region outline.

The present disclosure generally pertains to image sensor control circuitry for event-based controlling of an image sensor, the image sensor control circuitry being configured to: obtain events from a plurality of event-based vision elements of an event-based vision sensor; determine event groups based on an event-detection property; and generate an imaging control signal for controlling the imaging elements of the image sensor based on the event groups, for imaging with imaging element groups corresponding to the event groups.

The present disclosure generally pertains to image sensor control circuitry for event-based controlling of an image sensor, the image sensor control circuitry being configured to: obtain events from a plurality of event-based vision elements of an event-based vision sensor; determine event groups based on an event-detection property; and generate an imaging control signal for controlling the imaging elements of the image sensor based on the event groups, for imaging with imaging element groups corresponding to the event groups.

A method for defending against an adversarial sample in image classification includes: denoising, by an adversarial denoising network, an input image to acquire a reconstructed image; acquiring, by a target classification model, a predicted category probability distribution of the reconstructed image; acquiring, by the target classification model, a predicted category probability distribution of the original input image; calculating an adversarial score of the input image, and determining the input image as an adversarial sample or a benign sample according to a threshold; outputting a category prediction result of the reconstructed image if the input image is determined as the adversarial sample; and outputting a category prediction result of the original input image if the input image is determined as the benign sample. A system for defending against an adversarial sample in image classification, and a data processing terminal are further provided.

The current disclosure provides methods and systems to reduce an amount of structured and unstructured noise in image data. Specifically, a multi-stage deep learning method is provided, comprising training a deep learning network using a set of training pairs interchangeably including input data from a first noisy dataset with a first noise level and target data from a second noisy dataset with a second noise level, and input data from the second noisy dataset and target data from the first noisy dataset; generating an ultra-low noise data equivalent based on a low noise data fed into the trained deep learning network; and retraining the deep learning network on the set of training pairs using the target data of the set of training pairs in a first retraining step, and using the ultra-low noise data equivalent as target data in a second retraining step.

Described herein is a method for training a denoising model. The method includes obtaining a first set of simulated images based on design patterns. The simulated images may be clean and can be added with noise to generate noisy simulated images. The simulated clean and noisy images are used as training data to generate a denoising model.

An apparatus includes a memory and processor. The memory stores OCR and NLP algorithms. The processor receives an image of a physical document page and executes the OCR algorithm to convert the image into text. The processor identifies errors in the text, which are associated with noise in the image. The processor generates a feature vector that includes features obtained by executing the NLP algorithm on the text, and features associated with the identified errors in the text. The processor uses the feature vector to assign the image to a document category. Documents assigned to the document category share one or more characteristics, and the feature vector is associated with a probability greater than a threshold that the physical document associated with the image includes those characteristics. The processor then stores the image in a database as a page of an electronic document belonging to the assigned document category.

An apparatus includes a memory and processor. The memory stores OCR and NLP algorithms. The processor receives an image of a physical document page and executes the OCR algorithm to convert the image into text. The processor identifies errors in the text, which are associated with noise in the image. The processor generates a feature vector that includes features obtained by executing the NLP algorithm on the text, and features associated with the identified errors in the text. The processor uses the feature vector to assign the image to a document category. Documents assigned to the document category share one or more characteristics, and the feature vector is associated with a probability greater than a threshold that the physical document associated with the image includes those characteristics. The processor then stores the image in a database as a page of an electronic document belonging to the assigned document category.