no code implementations • 24 Jun 2023 • Hidenori Itaya, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi, Komei Sugiura
The decoder in AQT utilizes action queries, which represent the information of each action, as queries.
no code implementations • 4 Jun 2023 • Kohei Hattori, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi
In this study, based on a deep-learning model that incorporates the knowledge of experts, a method by which a learner "learns from AI" the grounds for its decisions is proposed.
no code implementations • 13 Apr 2023 • Yucheng Zhang, Masaki Fukuda, Yasunori Ishii, Kyoko Ohshima, Takayoshi Yamashita
Unlike 2D image labels, annotating point cloud data is difficult due to the limitations of sparsity, irregularity, and low resolution, which requires more manual work, and the annotation efficiency is much lower than 2D image. Therefore, we propose an annotation algorithm for point cloud data, which is pre-annotation and camera-LiDAR late fusion algorithm to easily and accurately annotate.
no code implementations • 30 Mar 2023 • Nobuhiko Wakai, Satoshi Sato, Yasunori Ishii, Takayoshi Yamashita
In orthogonal world coordinates, a Manhattan world lying along cuboid buildings is widely useful for various computer vision tasks.
no code implementations • 16 Feb 2023 • Hiroki Adachi, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi, Yasunori Ishii, Kazuki Kozuka
Adversarial training is a popular and straightforward technique to defend against the threat of adversarial examples.
no code implementations • 12 Sep 2022 • Shungo Fujii, Yasunori Ishii, Kazuki Kozuka, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi
Data augmentation is an essential technique for improving recognition accuracy in object recognition using deep learning.
no code implementations • 31 Aug 2022 • Yuzuru Nakamura, Yasunori Ishii, Yuki Maruyama, Takayoshi Yamashita
In object detection, data amount and cost are a trade-off, and collecting a large amount of data in a specific domain is labor intensive.
no code implementations • 11 May 2022 • Risako Tanigawa, Yasunori Ishii, Kazuki Kozuka, Takayoshi Yamashita
They are highly and widely used in tasks such as segmentation.
no code implementations • 15 Apr 2022 • Risako Tanigawa, Yasunori Ishii, Kazuki Kozuka, Takayoshi Yamashita
In inference, it is possible to obtain instance segmentation results only from sound images.
no code implementations • 25 Nov 2021 • Nobuhiko Wakai, Satoshi Sato, Yasunori Ishii, Takayoshi Yamashita
To address this problem, we propose a generic camera model that has the potential to address various types of distortion.
1 code implementation • 29 Oct 2021 • Masahiro Mitsuhara, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi
It is difficult for people to interpret the decision-making in the inference process of deep neural networks.
no code implementations • 16 Jul 2021 • Yasunori Ishii, Takayoshi Yamashita
It is difficult to collect data on a large scale in a monocular depth estimation because the task requires the simultaneous acquisition of RGB images and depths.
Ranked #42 on Monocular Depth Estimation on NYU-Depth V2 (using extra training data)
no code implementations • 27 Mar 2021 • Naoki Okamoto, Soma Minami, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi
In this study, we propose an ensemble method using knowledge transfer to improve the accuracy of ensembles by introducing a loss design that promotes diversity among networks in mutual learning.
no code implementations • 6 Mar 2021 • Hidenori Itaya, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi, Komei Sugiura
A3C consists of a feature extractor that extracts features from an image, a policy branch that outputs the policy, and a value branch that outputs the state value.
no code implementations • 12 Feb 2021 • Aly Magassouba, Komei Sugiura, Angelica Nakayama, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi, Hisashi Kawai
Thus, inferring the collision-risk before a placing motion is crucial for achieving the requested task.
no code implementations • 9 Jul 2020 • Tadashi Ogura, Aly Magassouba, Komei Sugiura, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi, Hisashi Kawai
Domestic service robots (DSRs) are a promising solution to the shortage of home care workers.
1 code implementation • 10 Sep 2019 • Soma Minami, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi
To achieve the knowledge transfer, we propose a novel graph representation called knowledge transfer graph that provides a unified view of the knowledge transfer and has the potential to represent diverse knowledge transfer patterns.
no code implementations • 9 May 2019 • Masahiro Mitsuhara, Hiroshi Fukui, Yusuke Sakashita, Takanori Ogata, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi
As a result, the fine-tuned network can output an attention map that takes into account human knowledge.
3 code implementations • CVPR 2019 • Hiroshi Fukui, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi
ABN can be applicable to several image recognition tasks by introducing a branch for attention mechanism and is trainable for the visual explanation and image recognition in end-to-end manner.
no code implementations • 1 Nov 2018 • Tsubasa Hirakawa, Takayoshi Yamashita, Toru Tamaki, Hironobu Fujiyoshi
Because path prediction as a task of computer vision uses video as input, various information used for prediction, such as the environment surrounding the target and the internal state of the target, need to be estimated from the video in addition to predicting paths.
no code implementations • 30 Oct 2017 • Masaya Hibino, Akisato Kimura, Takayoshi Yamashita, Yuji Yamauchi, Hironobu Fujiyoshi
A denoising autoencoder can be trained with indicator vectors produced from clean and noisy input samples, and non-leaf nodes where incorrect decisions are made can be identified by comparing the input and output of the trained denoising autoencoder.
no code implementations • 14 Sep 2017 • Ryuji Kamiya, Takayoshi Yamashita, Mitsuru Ambai, Ikuro Sato, Yuji Yamauchi, Hironobu Fujiyoshi
Our method replaces real-valued inner-product computations with binary inner-product computations in existing network models to accelerate computation of inference and decrease model size without the need for retraining.
no code implementations • ICCV 2015 • Takahiro Hasegawa, Mitsuru Ambai, Kohta Ishikawa, Gou Koutaki, Yuji Yamauchi, Takayoshi Yamashita, Hironobu Fujiyoshi
We propose a method for estimating multiple-hypothesis affine regions from a keypoint by using an anisotropic Laplacian-of-Gaussian (LoG) filter.