The disclosure relates to a system and a method for generating a neural network model for image processing by interacting with at least one client terminal. The method may include receiving via a network, a plurality of first training samples from the at least one client terminal. The method may also include training a first neural network model based on the plurality of first training samples to generate a second neural network model. The method may further include transmitting, via the network, the second neural network model to the at least one client terminal.

The disclosure relates to a system and a method for generating a neural network model for image processing by interacting with at least one client terminal. The method may include receiving via a network, a plurality of first training samples from the at least one client terminal. The method may also include training a first neural network model based on the plurality of first training samples to generate a second neural network model. The method may further include transmitting, via the network, the second neural network model to the at least one client terminal.

Aspects of the disclosure provide an image processing method, an image processing terminal, an image processing apparatus, and a non-transitory computer-readable storage medium. The method can include performing face detection on a target image to detect a face region in the target image. The method includes determining an expression class of the detected face region based on a trained expression recognition model and determining whether the determined expression class of the detected face region is a target expression class. If the determined expression class is the target expression class, head detection is performed to identify a head region in the target image and the identified head region is edited. If the determined expression class is not the target expression class, the head detection is not performed on the target image.

Aspects of the disclosure provide an image processing method, an image processing terminal, an image processing apparatus, and a non-transitory computer-readable storage medium. The method can include performing face detection on a target image to detect a face region in the target image. The method includes determining an expression class of the detected face region based on a trained expression recognition model and determining whether the determined expression class of the detected face region is a target expression class. If the determined expression class is the target expression class, head detection is performed to identify a head region in the target image and the identified head region is edited. If the determined expression class is not the target expression class, the head detection is not performed on the target image.

A method for recognizing which lamps at a vehicle are activated. The method includes: providing multiple image recordings of candidate areas at the vehicle in which an activated lamp is presumed; converting the image recordings into an intermediate product by executing a recurrent encoder network (ERNN), the output of at least one pass of the ERNN is supplied as input to a further pass of the ERNN, and different image recordings of candidate areas are supplied as input to different passes of the ERNN; assignments of the image recordings of candidate areas are ascertained to classes which represent specific lamps of the vehicle from the intermediate product by executing a recurrent decoder network (DRNN) multiple times, the output of at least one pass of the DRNN is supplied as input to a further pass of the DRNN.

A method for recognizing which lamps at a vehicle are activated. The method includes: providing multiple image recordings of candidate areas at the vehicle in which an activated lamp is presumed; converting the image recordings into an intermediate product by executing a recurrent encoder network (ERNN), the output of at least one pass of the ERNN is supplied as input to a further pass of the ERNN, and different image recordings of candidate areas are supplied as input to different passes of the ERNN; assignments of the image recordings of candidate areas are ascertained to classes which represent specific lamps of the vehicle from the intermediate product by executing a recurrent decoder network (DRNN) multiple times, the output of at least one pass of the DRNN is supplied as input to a further pass of the DRNN.

A method is provided for identifying consumer packaged goods (CPGs). The method includes providing to a machine learning classifier a set of images containing at least one CPG; receiving from the machine learning classifier an indication that the machine learning classifier cannot reliably identify a designated CPG in the set of images; determining whether the designated CPG is a product in a product catalog; if the designated CPG is in the product catalog, then associating the designated CPG with a Global Trade Item Number (GTIN); and if the designated product is not in the product catalog, then designating the CPG as a potentially new product. Notably, this approach allows partially identified products to be treated as full-fledged members of the product catalog, thus allowing data to be collected on these products even before they have been fully identified and their GTINs have been resolved.

A method is provided for identifying consumer packaged goods (CPGs). The method includes providing to a machine learning classifier a set of images containing at least one CPG; receiving from the machine learning classifier an indication that the machine learning classifier cannot reliably identify a designated CPG in the set of images; determining whether the designated CPG is a product in a product catalog; if the designated CPG is in the product catalog, then associating the designated CPG with a Global Trade Item Number (GTIN); and if the designated product is not in the product catalog, then designating the CPG as a potentially new product. Notably, this approach allows partially identified products to be treated as full-fledged members of the product catalog, thus allowing data to be collected on these products even before they have been fully identified and their GTINs have been resolved.

Methods, systems, articles of manufacture and apparatus are disclosed to label text on images. An example apparatus includes colorizer circuitry to apply color to text boxes corresponding to optical character recognition (OCR) data associated with an image, OCR manager circuitry to render an OCR text prompt associated with the OCR data, the OCR text prompt to be rendered proximate to respective ones of the text boxes, the OCR text prompt to display a text portion of the OCR data, and edit circuitry to (a) render an interface in response to selection of the OCR text prompt, the interface populated with the text portion of the OCR data, and (b) in response to an overwrite input to the interface, update the text portion of the OCR data in a memory corresponding to the image.

A device includes image generation circuitry and convolutional-neural-network circuitry. The image generation circuitry, in operation, generates a digital image representation of a wafer defect map (WDM). The convolutional-neural-network circuitry, in operation, generates a defect classification associated with the WDM based on: the digital image representation of the WDM and a data-driven model associating WDM images with classes of a defined set of classes of wafer defects and generated using a training data set augmented based on defect pattern orientation types associated with training images.