Abstract: An apparatus for calibrating confidence determines a correct logit element and a false logit element among a plurality of logit elements of a logit vector obtained by a neural network model performing a neural network operation on the training data, based on a correct class and a false class determined from a target probability distribution including target probabilities for each of a plurality of classes according to ground truth of the training data as elements, and determines whether to calibrate and a calibration value for a plurality of target probabilities of the target probability distribution based on a difference in value between the correct logit element and the false logit element.

Abstract: A device, and method thereof, for training a model for human identification are provided may include a primary training module configured to primarily train the model with respect to a pre-prepared source dataset, a target subset generation module configured to generate a target subset by selecting some cameras from among a plurality of cameras included on a service robot, a feature vector extraction module configured to extract feature vectors of the target subset by using the model, a labeling module configured to perform labeling on the feature vectors, and a secondary training module configured to secondarily train the model with respect to a target dataset by using results of the labeling.

Abstract: A method for searching a neural network architecture using supernets comprises the steps of: (a) searching for subnets that can be extracted from a set search space; (b) counting the number of non-linear activation functions included in each subnet for each of the searched subnets; (c) grouping the searched subnets based on the counted number of non-linear activation functions; (d) assigning the subnet groups to multiple supernets; (e) searching for a neural network having an optimal architecture based on operation blocks of the subnet groups assigned to each of the multiple supernets.

Abstract: The present disclosure provides a quantization apparatus and method, and an image processing apparatus, capable of improving operation efficiency while suppressing performance degradation due to quantization errors by receiving one of a plurality of matrices obtained for neural network operation in an artificial neural network as an input matrix, dividing the input matrix into a plurality of channels, selecting a quantizer for each channel by analyzing the distribution of element values included in each of the plurality of divided channels, and quantizing element values included in the channels using the selected quantizer.

Abstract: A person re-identification device comprises: a feature extracting and dividing unit, that receives images including a person to be re-identified and extracts a feature of each image according to a pre-learned pattern estimation method to acquire a 3-dimensional feature vector, and divides the 3-dimensional feature vector into a pre-designated size unit to acquire local feature vectors; a one-to-many relational reasoning unit, that estimates the relationship between each of the local feature vectors and remaining local feature vectors, and reflects the estimated relationship to acquire local relational features; a global contrastive pooling unit, that acquires a global contrastive feature by performing global contrastive pooling; and a person re-identification unit, that receives the local relational features and the global contrastive feature as a final descriptor of a corresponding image, and compares the final descriptor with a reference descriptor acquired in advance from an image including a person to be

Abstract: Provided is an operation method of a computing device for providing a confidence calibration function based on label smoothing. The operation method includes inputting image data to be classified into an image classification model, outputting logit output values and prediction probability values of the image data by the image classification model, and outputting a label smoothing loss function by the image classification model by using learnable margin parameters or a label smoothing weighting based on an image classification difficulty level or all.

Abstract: In an embodiment an apparatus includes a processor configured to generate a feature extraction module using a dataset in which an attribute for each object is defined, receive an image obtained by a camera, extract an attribute of an object of interest from the image using the learned feature extraction module, identify an object re-identification candidate group based on the extracted attribute of the object of interest and re-identify the object of interest based on the identified object re-identification candidate group.

Abstract: An apparatus and a method for re-identifying an object to improve object re-identification performance using attribute information of the object are provided. The apparatus trains an object representation extraction model to train attribute representations, inputs an image obtained by a camera to the trained object representation extraction model, extracts object representations from the image using the trained object representation extraction model, and performs object re-identification based on the object representations.

Abstract: A person re-identification device comprises: a feature extracting and dividing unit, that receives images including a person to be re-identified and extracts a feature of each image according to a pre-learned pattern estimation method to acquire a 3-dimensional feature vector, and divides the 3-dimensional feature vector into a pre-designated size unit to acquire local feature vectors; a one-to-many relational reasoning unit, that estimates the relationship between each of the local feature vectors and remaining local feature vectors, and reflects the estimated relationship to acquire local relational features; a global contrastive pooling unit, that acquires a global contrastive feature by performing global contrastive pooling; and a person re-identification unit, that receives the local relational features and the global contrastive feature as a final descriptor of a corresponding image, and compares the final descriptor with a reference descriptor acquired in advance from an image including a person to be

Abstract: The present disclosure can provide a person re-identification apparatus and method that includes an identity feature extraction part configured to receive a multiple number of images each including a person requiring re-identification, extract features related to an identity of a person included in each image according to a pattern estimation method learned beforehand, and obtain an identity-related feature vector for each image; and a re-identification determination part configured to analyze a degree of similarity between an identity-related feature vector obtained for a base image including a search target from among the plurality of images and an identity-related feature vector obtained for another image to determine whether or not a person corresponding to the search target is included in the other image.

Abstract: The present disclosure can provide a person re-identification apparatus and method that includes an identity feature extraction part configured to receive a multiple number of images each including a person requiring re-identification, extract features related to an identity of a person included in each image according to a pattern estimation method learned beforehand, and obtain an identity-related feature vector for each image; and a re-identification determination part configured to analyze a degree of similarity between an identity-related feature vector obtained for a base image including a search target from among the plurality of images and an identity-related feature vector obtained for another image to determine whether or not a person corresponding to the search target is included in the other image.

Abstract: A content retargeting method and apparatus are disclosed. An embodiment of the invention provides a content retargeting method that includes: dividing an original content into a grid having M rows×N columns of quads; computing degrees of importance of the divided quads; and scaling the quads based on the computed degrees of importance, the quads axis-aligned by rows or columns.

Abstract: A content retargeting method and apparatus are disclosed. An embodiment of the invention provides a content retargeting method that includes: dividing an original content into a grid having M rows×N columns of quads; computing degrees of importance of the divided quads; and scaling the quads based on the computed degrees of importance, the quads axis-aligned by rows or columns.

Abstract: Disclosed are an apparatus and method for synthesizing an intermediate view image and a recording medium thereof. The disclosed apparatus for synthesizing an intermediate view image may include: a probability information generating part configured to generate information related to a matching probability between a first image and a second image; and a synthesizing part configured to synthesize an intermediate view image by applying a weight according to the matching probability to the first image and second image. Certain embodiments of the invention provide the advantage of enabling enhanced image quality when synthesizing an intermediate view image without requiring post-processing operations such as for processing occlusions and for filling holes.