Akisato Kimura
YOU?
Author Swipe
View article: Second-timescale Glints from Satellites and Space Debris Detected with Tomo-e Gozen
Second-timescale Glints from Satellites and Space Debris Detected with Tomo-e Gozen Open
A search for second-timescale optical transients is one of the frontiers of time-domain astronomy. However, it has been pointed out that reflections of sunlight from Earth-orbiting objects can also produce second-timescale ``glints.'' We c…
View article: Neural network prediction of model parameters for strong lensing samples from Hyper Suprime-Cam Survey
Neural network prediction of model parameters for strong lensing samples from Hyper Suprime-Cam Survey Open
Strong lensing of background galaxies provides important information about the matter distribution around lens galaxies. Traditional modelling of such strong lenses is both time and resource intensive. Fast and automated analysis methods a…
View article: Personalization of Code Readability Evaluation Based on LLM Using Collaborative Filtering
Personalization of Code Readability Evaluation Based on LLM Using Collaborative Filtering Open
Code readability is an important indicator of software maintenance as it can significantly impact maintenance efforts. Recently, LLM (large language models) have been utilized for code readability evaluation. However, readability evaluatio…
View article: Acoustic-based 3D Human Pose Estimation Robust to Human Position
Acoustic-based 3D Human Pose Estimation Robust to Human Position Open
This paper explores the problem of 3D human pose estimation from only low-level acoustic signals. The existing active acoustic sensing-based approach for 3D human pose estimation implicitly assumes that the target user is positioned along …
View article: Neural network prediction of model parameters for strong lensing samples from Hyper Suprime-Cam Survey
Neural network prediction of model parameters for strong lensing samples from Hyper Suprime-Cam Survey Open
Strong lensing of background galaxies provides important information about the matter distribution around lens galaxies. Traditional modelling of such strong lenses is both time and resource intensive. Fast and automated analysis methods a…
View article: Unsupervised Intrinsic Image Decomposition with LiDAR Intensity Enhanced Training
Unsupervised Intrinsic Image Decomposition with LiDAR Intensity Enhanced Training Open
Unsupervised intrinsic image decomposition (IID) is the process of separating a natural image into albedo and shade without these ground truths. A recent model employing light detection and ranging (LiDAR) intensity demonstrated impressive…
View article: Deep Attentive Time Warping
Deep Attentive Time Warping Open
Similarity measures for time series are important problems for time series classification. To handle the nonlinear time distortions, Dynamic Time Warping (DTW) has been widely used. However, DTW is not learnable and suffers from a trade-of…
View article: Toward Defensive Letter Design
Toward Defensive Letter Design Open
A major approach for defending against adversarial attacks aims at controlling only image classifiers to be more resilient, and it does not care about visual objects, such as pandas and cars, in images. This means that visual objects thems…
View article: Selective Scene Text Removal
Selective Scene Text Removal Open
Scene text removal (STR) is the image transformation task to remove text regions in scene images. The conventional STR methods remove all scene text. This means that the existing methods cannot select text to be removed. In this paper, we …
View article: Deep Quantigraphic Image Enhancement via Comparametric Equations
Deep Quantigraphic Image Enhancement via Comparametric Equations Open
Most recent methods of deep image enhancement can be generally classified into two types: decompose-and-enhance and illumination estimation-centric. The former is usually less efficient, and the latter is constrained by a strong assumption…
View article: ConceptBeam
ConceptBeam Open
We propose a novel framework for target speech extraction based on semantic\ninformation, called ConceptBeam. Target speech extraction means extracting the\nspeech of a target speaker in a mixture. Typical approaches have been\nexploiting …
View article: Reflectance-Oriented Probabilistic Equalization for Image Enhancement
Reflectance-Oriented Probabilistic Equalization for Image Enhancement Open
Despite recent advances in image enhancement, it remains difficult for existing approaches to adaptively improve the brightness and contrast for both low-light and normal-light images. To solve this problem, we propose a novel 2D histogram…
View article: Font Generation with Missing Impression Labels
Font Generation with Missing Impression Labels Open
Our goal is to generate fonts with specific impressions, by training a generative adversarial network with a font dataset with impression labels. The main difficulty is that font impression is ambiguous and the absence of an impression lab…
View article: Font Shape-to-Impression Translation
Font Shape-to-Impression Translation Open
Different fonts have different impressions, such as elegant, scary, and cool. This paper tackles part-based shape-impression analysis based on the Transformer architecture, which is able to handle the correlation among local parts by its s…
View article: Prototyping of low-cost color enhancement lighting using multicolor LEDs
Prototyping of low-cost color enhancement lighting using multicolor LEDs Open
We prototyped a lighting system for color enhancement with maintaining a white appearance using low-cost multicolor LEDs. We bought LEDs of several colors in local electronic parts shops, and spectral power distributions of them were evalu…
View article: Color saturation control by modulating spectral power distribution of illumination using color enhancement factors
Color saturation control by modulating spectral power distribution of illumination using color enhancement factors Open
Color enhancement factors are spectral components that modulate the spectral power distribution (SPD) of illumination for interactive color saturation control. They can enhance more than one target color simultaneously or independently whi…
View article: Creating Stories: Social Curation of Twitter Messages
Creating Stories: Social Curation of Twitter Messages Open
Social media has become ubiquitous. Tweets and other user-generated content have become so abundant that better tools for information organization are needed in order to fully exploit their potential richness. ”Social cu- ration” has recen…
View article: Attention to Warp: Deep Metric Learning for Multivariate Time Series
Attention to Warp: Deep Metric Learning for Multivariate Time Series Open
Deep time series metric learning is challenging due to the difficult trade-off between temporal invariance to nonlinear distortion and discriminative power in identifying non-matching sequences. This paper proposes a novel neural network-b…
View article: Which Parts Determine the Impression of the Font?
Which Parts Determine the Impression of the Font? Open
Various fonts give different impressions, such as legible, rough, and comic-text.This paper aims to analyze the correlation between the local shapes, or parts, and the impression of fonts. By focusing on local shapes instead of the whole l…
View article: Shared Latent Space of Font Shapes and Their Noisy Impressions
Shared Latent Space of Font Shapes and Their Noisy Impressions Open
Styles of typefaces or fonts are often associated with specific impressions, such as heavy, contemporary, or elegant. This indicates that there are certain correlations between font shapes and their impressions. To understand the correlati…
View article: Shared Latent Space of Font Shapes and Impressions.
Shared Latent Space of Font Shapes and Impressions. Open
We have specific impressions from the style of a typeface (font), suggesting that there are correlations between font shape and its impressions. Based on this hypothesis, we realize a shared latent space where a font shape image and its im…
View article: Impressions2Font: Generating Fonts by Specifying Impressions
Impressions2Font: Generating Fonts by Specifying Impressions Open
Various fonts give us various impressions, which are often represented by words. This paper proposes Impressions2Font (Imp2Font) that generates font images with specific impressions. Imp2Font is an extended version of conditional generativ…
View article: Cross-media Scene Analysis: Estimating Objects' Visuals Only from Audio
Cross-media Scene Analysis: Estimating Objects' Visuals Only from Audio Open
Human beings can get a visual image of the surrounding environment from sounds they hear.Can we give similar capabilities to computers?In this article, we introduce our recent efforts in cross-media scene analysis applied to estimate the t…
View article: Weakly Supervised Collective Feature Learning From Curated Media
Weakly Supervised Collective Feature Learning From Curated Media Open
The current state-of-the-art in feature learning relies on the supervised learning of large-scale datasets consisting of target content items and their respective category labels. However, constructing such large-scale fully-labeled datase…
View article: Weakly supervised collective feature learning from curated media
Weakly supervised collective feature learning from curated media Open
The current state-of-the-art in feature learning relies on the supervised learning of large-scale datasets consisting of target content items and their respective category labels. However, constructing such large-scale fully-labeled datase…
View article: Few-shot learning of neural networks from scratch by pseudo example optimization
Few-shot learning of neural networks from scratch by pseudo example optimization Open
In this paper, we propose a simple but effective method for training neural networks with a limited amount of training data. Our approach inherits the idea of knowledge distillation that transfers knowledge from a deep or wide reference mo…
View article: Imitation networks: Few-shot learning of neural networks from scratch
Imitation networks: Few-shot learning of neural networks from scratch Open
In this paper, we propose imitation networks, a simple but effective method for training neural networks with a limited amount of training data. Our approach inherits the idea of knowledge distillation that transfers knowledge from a deep …
View article: Weakly Supervised Collective Feature Learning from Curated Media
Weakly Supervised Collective Feature Learning from Curated Media Open
The current state-of-the-art in feature learning relies on the supervised learning of large-scale datasets consisting of target content items and their respective category labels. However, constructing such large-scale fully-labeled datase…
View article: Denoising random forests
Denoising random forests Open
This paper proposes a novel type of random forests called a denoising random forests that are robust against noises contained in test samples. Such noise-corrupted samples cause serious damage to the estimation performances of random fores…