Methods, apparatuses and computer program products for providing artificial-intelligence-based indicia data editing are provided. For example, an example computer-implemented method may include determining, based at least in part on a data processing model associated with a scan setting module, a first decoded data string corresponding to a first indicia; determining, based at least in part on user input data, a first input data string corresponding to the first indicia; generating a predictive indicia data editing model based at least in part on providing the first decoded data string and the first input data string to an artificial intelligence algorithm; and updating the scan setting module based at least in part on the predictive indicia data editing model.

A method for processing an encoded pattern, a storage medium, and an electronic device are disclosed in this application. An electronic device acquires a first encoded pattern to be recognized, the first encoded pattern being a pattern obtained after predefined information is encoded therein. The electronic device increases resolution of the first encoded pattern through a target model to obtain a second encoded pattern, the target model being obtained through training by using a predetermined third encoded pattern as an input and a predetermined fourth encoded pattern as an output, the third encoded pattern being obtained by decreasing resolution of the fourth encoded pattern, the third and the first encoded patterns being encoded in the same manner. Finally, the electronic device decodes the second encoded pattern using a code recognition module to obtain the predefined information. This application resolves the technical problem that an encoded pattern cannot be accurately recognized.

A device receives image data that depicts a target that is subject to a security check, and receives transactional data identifying a characteristic of a transaction associated with the target. The device identifies the target within the image data and identifies a first set of target attributes of the target. The device determines a risk level that represents a likelihood of the characteristic of the transaction having a correct value by analyzing the set of target attributes using one or more attribute recognition techniques and/or using a data model that has been trained using machine learning. The device determines whether the risk level satisfies one or more threshold risk levels. The device provides an alert to another device cause the other device to perform actions based on whether at least one of the one or more threshold risk levels are satisfied.

A device receives image data that depicts an individual, identifies the individual by using a target recognition technique to analyze the image data, and identifies human attributes of the individual. The device determines a risk level that represents a likelihood of the individual being or becoming involved in a harmful act. The risk level may be determined by using the human attributes as part of a machine-learning-driven emotional analysis of the individual. The device provides a message, selected based on whether the risk level satisfies at least one of the one or more threshold risk levels, to another device, to cause the other device to perform actions that are indicative of the risk level being associated with a low amount of risk, or other actions indicative of the risk level being associated with a high amount of risk and that are associated with assisting in prevention of the harmful act.

Techniques are provided for detecting an improper barcode using a neural network trained to identify an object from physical features appearing in images of object, and without resorting to using a barcode or other indicia to identify the object. The neural network is self-training, updating itself with selected images obtained at a Point-of-Sale. That is, the neural network is capable of training itself while performing improper barcode detection operations, such as spoofing detection.

Techniques are provided for training a neural network, where the techniques include receiving image scan data of an object, such as a product or package presented at a scanning station, where the image scan data includes an image that contains at least one indicia corresponding to the object and physical features of the object. A neural networks examines the physical features and determines weighting indicating a correlation strength between the physical feature and an identification data of the object. Thereby training a neural network to identify objects from their scanned physical features in place or in accompaniment to scanned indicia data.

Techniques are provided for detecting an improper barcode using a neural network trained to identify an object from physical features appearing in images of object, and without resorting to using a barcode or other indicia to identify the object. The neural network is self-training, updating itself with selected images obtained at a Point-of-Sale. That is, the neural network is capable of training itself while performing improper barcode detection operations, such as spoofing detection.

Techniques for improving the accuracy of a neural network trained for loss prevention applications include identifying physical features of an object in image scan data, cropping indicia from the image scan data, and examining physical features in the indicia-removed image scan data using a neural network to identify the object based on comparison of identification data based on the physical features and other identification, such as based on the indicia. In response to a match prediction, indicating a match and generating an authenticating signal.

Systems and methods for establishing optimal reading conditions for a machine-readable symbol reader. A machine-readable symbol reader may selectively control reading conditions including lighting conditions (e.g., illumination pattern), focus, decoder library parameters (e.g., exposure time, gain), etc. Deep learning and optimization algorithms (e.g., greedy search algorithms) are used to autonomously learn an optimal set of reading parameters to be used for the reader in a particular application. A deep learning network (e.g., a convolutional neural network) may be used to locate machine-readable symbols in images captured by the reader, and greedy search algorithms may be used to determine a reading distance parameter and one or more illumination parameters during an autonomous learning phase of the reader. The machine-readable symbol reader may be configured with the autonomously learned reading parameters, which enables the machine-readable symbol reader to accurately and quickly decode machine-readable symbols (e.g., direct part marking (DPM) symbols).

A device receives image data that depicts an individual, identifies the individual by using a target recognition technique to analyze the image data, and identifies human attributes of the individual. The device determines a risk level that represents a likelihood of the individual being or becoming involved in a harmful act. The risk level may be determined by using the human attributes as part of a machine-learning-driven emotional analysis of the individual. The device provides a message, selected based on whether the risk level satisfies at least one of the one or more threshold risk levels, to another device, to cause the other device to perform actions that are indicative of the risk level being associated with a low amount of risk, or other actions indicative of the risk level being associated with a high amount of risk and that are associated with assisting in prevention of the harmful act.