Abstract:
A method and apparatus with object detector training is provided. The method includes obtaining first input data and second input data from a target object; obtaining second additional input data by performing data augmentation on the second input data; extracting a first feature to a shared embedding space by inputting the first input data to a first encoder; extracting a second feature to the shared embedding space by inputting the second input data to a second encoder; extracting a second additional feature to the shared embedding space by inputting thesecond additional input data to the second encoder; identifying a first loss function based on the first feature, the second feature, and the second additional feature; identifying a second loss function based on the second feature and the second additional feature; and updating a weight of the second encoder based on the first loss function and the second loss function.
Abstract:
A method of generating a hyperlapse video includes: comparing a first reference point of a first image and a corresponding second reference point of a second image; based on the comparing, displaying a first user interface for matching the first reference point and second reference point; and determining whether to perform automatic shooting for the hyperlapse video based on whether the first reference point and the second reference point match.
Abstract:
A processor-implemented method with sensor calibration includes: estimating a portion of a rotation parameter for a target sensor among a plurality of sensors based on a capture of a reference object; estimating another portion of the rotation parameter for the target sensor based on an intrinsic parameter of the target sensor and a focus of expansion (FOE) determined based on sensing data collected with consecutive frames by the target sensor while the electronic device rectilinearly moves based on one axis; and performing calibration by determining a first extrinsic parameter for the target sensor based on the portion and the other portion of the rotation parameter.
Abstract:
Disclosed is a method and apparatus for adaptive tracking of a target object. The method includes method of tracking an object, the method including estimating a dynamic characteristic of an object in an input image based on frames of the input image, determining a size of a crop region for a current frame of the input image based on the dynamic characteristic of the object, generating a cropped image by cropping the current frame based on the size of the crop region, and generating a result of tracking the object for the current frame using the cropped image.
Abstract:
A processor-implemented neural network method includes: obtaining a first weight kernel of a weight model and pruning information of the first weight kernel; determining, based on the pruning information, a processing range of an input feature map for each weight element vector of the first weight kernel; performing a convolution operation between the input feature map and the first weight kernel based on the determined processing range; and generating an output feature map of a neural network layer based on an operation result of the convolution operation.
Abstract:
Provided are an optical sensor including graphene quantum dots and an image sensor including an optical sensing layer. The optical sensor may include a graphene quantum dot layer that includes a plurality of first graphene quantum dots bonded to a first functional group and a plurality of second graphene quantum dots bonded to a second functional group that is different from the first functional group. An absorption wavelength band of the optical sensor may be adjusted based on types of functional groups bonded to the respective graphene quantum dots and/or sizes of the graphene quantum dots.
Abstract:
Disclosed are heat dissipation structures using nano-sized graphene fragments such as graphene quantum dots (GQDs) and/or methods of manufacturing the heat dissipation structures. A heat dissipation structure includes a heating element, and a heat dissipation film on the heating element to dissipate heat generated from the heating element, to outside. The heat dissipation film may include GQDs.
Abstract:
A method of manufacturing an organic-inorganic composite thin film may include: forming a thin film from a paste that includes an inorganic powder and an organic compound binder by using a screen printing process; and/or performing a pressing process and a heating process with respect to the thin film. The heating process may be performed at a glass transition temperature of the organic compound binder or in a temperature range higher than the glass transition temperature of the organic compound binder. An X-ray detector configured to detect X-rays irradiated from an outside of the X-ray detector may include: a photoconductive material layer in which electron-hole pairs are formed due to absorption of the X-rays. The photoconductive material layer may be formed of an organic-inorganic composite thin film that includes an inorganic powder and an organic compound binder.
Abstract:
Methods and systems for dynamically updating firmware of a single-protocol based one-chip radio device present in an Internet of things (IoT) environment with multiple technologies using an intelligent firmware update based on size of the flash memory, the one or more hardware resources available at the controller level, and a current IoT context associated with the IoT environment are provided. The method includes receiving a firmware update package, determining a size of flash memory and one or more hardware resources at a controller level of the single-protocol, based on receiving the firmware update package, correlating the received firmware update package with the size of the flash memory, dynamically selecting one or more firmware resources from the received firmware package for updating firmware of the single-protocol based one-chip radio device based on the correlation, and updating firmware of the single-protocol based one-chip radio device using the one or more firmware resources.
Abstract:
An apparatus includes one or more processors configured to generate a plurality of feature maps having respective different resolutions based on an input image; and update, for each of the plurality of transformer layers, respective position estimation information comprising first position information of a respective bounding box corresponding to one object query and second position information of respective key points corresponding to the one object query, wherein each of the plurality of transformer layers includes a self-attention model configured to generate respective intermediate data by performing self-attention on respective content information on a feature of the input image; and a cross-attention model configured to generate respective output data by performing cross-attention on respective one or more feature maps among the plurality of feature maps and the respective generated intermediate data.