Younghwan Chae

Improved Accuracy of Track-Based Integration of Radar and Vision Sensors

Seungheon Kwak, Younghwan Chae, Minyoung Choi, Hae-Seung Lim et autres

This paper proposes a method for integrating the camera and radar datas using non-linear homography transformations. The lens distortion effects of a camera cause differences in object positions to have a non-linear relationship with pixel differences in the captured image. Therefore, accurate …

kr (code pays fourni par la source)

Deep-Learning-Based Kick Motion Recognition in Millimeter Waveband Radar System

Chanul Park, Hyo-In Baek, Younghwan Chae, Hae-Seung Lim et autres

In this article, we propose a method for recognizing kick motions using a multiple-input multiple-output (MIMO) frequency-modulated continuous wave (FMCW) radar system combined with deep learning techniques. Smart trunk opener (STO) systems that provide users with hands-free trunk operation have been gaining …

kr (code pays fourni par la source)

Implementation of Deep Learning-based Kick Gesture Recognition Using 60 GHz Radar Sensor

Hyo-In Baek, Younghwan Chae, Hae-Seung Lim, Jae-Eun Lee et autres

In the realm of automotive technology, hands-free trunk operation systems have emerged as a cornerstone of convenience and functionality. Predominantly reliant on ultra-sonic and camera-based sensors, these systems, however, falter in adverse weather conditions such as rain, snow, and fog, and are …

kr (code pays fourni par la source)

Gradient-only surrogate to resolve learning rates for robust and consistent training of deep neural networks

Younghwan Chae, Daniel Nicolas Wilke, Dominic Kafka

Afrique du Sud (code pays fourni par la source)

GOALS: Gradient-Only Approximations for Line Searches Towards Robust and Consistent Training of Deep Neural Networks

Younghwan Chae, Daniel Nicolas Wilke, Dominic Kafka

Mini-batch sub-sampling (MBSS) is favored in deep neural network training to reduce the computational cost. Still, it introduces an inherent sampling error, making the selection of appropriate learning rates challenging. The sampling errors can manifest either as a bias or variances in …

Afrique du Sud (code pays fourni par la source)

Empirical study towards understanding line search approximations for training neural networks

Younghwan Chae, Daniel Nicolas Wilke

Choosing appropriate step sizes is critical for reducing the computational cost of training large-scale neural network models. Mini-batch sub-sampling (MBSS) is often employed for computational tractability. However, MBSS introduces a sampling error, that can manifest as a bias or variance in a …

Heuristic linear algebraic rank-variance formulation and solution approach for efficient sensor placement

Younghwan Chae, Daniel Nicolas Wilke

Afrique du Sud (code pays fourni par la source)

Optimal sensor placement approaches for the design of inverse experiments by simulation

Younghwan Chae