Embodiments relate to a method for re-identifying a target object based on location information of closed-circuit television (CCTV) and movement information of the target object and a system for performing the same, the method including detecting at least one object of interest in a plurality of source videos based on a preset condition of the object of interest, tracking the identified object of interest on the corresponding source video to generate a tube of the object of interest, receiving an image query including a target patch and location information of the CCTV, determining at least one search candidate area based on the location information of the CCTV and the movement information of the target object, re-identifying if the object of interest seen in the tube of the object of interest is the target object, and providing a user with the tube of the re-identified object of interest.

Embodiments of the present invention are directed to learning of an appropriate dictionary which has a high expression ability of minority data while preventing reduction of an expression ability of majority data. A dictionary generation apparatus which generates a dictionary used for discriminating whether data to be discriminated belongs to a specific category includes a generation unit configured to generate a first dictionary based on learning data belonging to the specific category and a selection unit configured to estimate a degree of matching of the learning data at each portion with the first dictionary using the generated first dictionary and select a portion of the learning data based on the estimated degree of matching, wherein the generation unit generates a second dictionary based on the selected portion of the learning data.

Provided is a configuration for generating pseudo sensor data from a plurality of pieces of existing sensor data. This information processing device which generates time-series learning data on the basis of time-series original data acquired from a robot device comprises: a memory that stores at least one extended data generation rule comprising at least one velocity change value, at least one phase change value, at least one position change value, or at least one magnitude change value; and a processor that generates time-series extended data by data expansion of the original data using at least one change value of the extended data generation rule, and outputs time-series learning data including the time-series extended data and the time-series original data.

A problem of supervised learning is overcome by using patches to discover objects in unlabeled training images. The discovered objects are embedded in a pattern space. An AI machine replaces manual entry steps of training with a machine-centric process including clustering in a pixel space, clustering in latent space and building the pattern space based on different losses derived from pixel space clustering and latent space clustering. A distance structure in the pattern space captures the co-occurrence of patterns due to frequently appearing objects in training image data. Embodiments provide image representation based on local image patch naturally handles the position and scale invariance property that is important to effective object detection. Embodiments successfully identifies frequent objects such as human faces, human bodies, animals, or vehicles from unorganized data images based on a small quantity of training images.

Techniques are provided for evaluating and defining the scope of data-driven deep learning models. In one embodiment, a machine-readable storage medium is provided comprising executable instructions that, when executed by a processor, facilitate performance of operations comprising employing a machine learning model to extract first training data features included in a training data set and first target data features included in a target data set. The operations further comprise determining whether the target data set is within a defined data scope of the training data set based on analysis of correspondences between the first training data features and the first target data feature, and determining whether application of the target data set to a target neural network model developed using the training data set will generate results with an acceptable level of accuracy based on whether the target data set is within the defined data scope.

Methods and systems are described herein for improvements to authenticate users, particularly authenticating a user based on data known to the user. For example, methods and systems allow for users to be securely authenticated based on data known to the users over remote communication networks without storing the data known to the users. Specifically, methods and systems authenticate users by requiring users to select images that are known to the users. For example, the methods and systems may generate synthetic images based on the user's own images and require the user to select the synthetic image, from a set of a set of images, that is known to the user to authenticate the user. Moreover, the methods and systems alleviate storage and privacy concerns by not storing the data known to the users.

An information processing apparatus selects, as a reference defect, at least one defect from among first defects associated with a first image and selects, as a correction target defect, at least one defect from among second defects associated with a second image captured at a time different from an image capturing time of the first image. Additionally, the information processing apparatus generates a correction candidate by modifying the correction target defect, acquires a matching level representing a matching relationship between the reference defect and the correction candidate, and generates a corrected defect by correcting the correction target defect based on the matching level. Then, the information processing apparatus acquires a progress level representing a change in defect from the reference defect based on a comparison between the reference defect and the corrected defect.

A neural network model receives color data for a sequence of images corresponding to a dynamic scene in three-dimensional (3D) space. Motion of objects in the image sequence results from a combination of a dynamic camera orientation and motion or a change in the shape of an object in the 3D space. The neural network model generates two components that are used to produce a 3D motion field representing the dynamic (non-rigid) part of the scene. The two components are information identifying dynamic and static portions of each image and the camera orientation. The dynamic portions of each image contain motion in the 3D space that is independent of the camera orientation. In other words, the motion in the 3D space (estimated 3D scene flow data) is separated from the motion of the camera.

A neural network is trained to identify one or more features of an image. The neural network is trained using a small number of original images, from which a plurality of additional images are derived. The additional images generated by rotating and decoding embeddings of the image in a latent space generated by an autoencoder. The images generated by the rotation and decoding exhibit changes to a feature that is in proportion to the amount of rotation.

An image analysis device 10 includes: a generation unit 11 which generates a similar set, which is a set of similar pieces of learning data selected from among a plurality of pieces of learning data, each including an image and information that represents an object to be recognized that is displayed in the image; and a learning unit 12 which uses the generated similar set to learn parameters for a predetermined recognition model that allow the predetermined recognition model to recognize the object to be recognized that is displayed in each image included in the generated similar set.