TAICHI YOSHIDA
| Department of Computer and Network Engineering | Associate Professor |
| Cluster II (Emerging Multi-interdisciplinary Engineering) | Associate Professor |
Researcher Information
Degree
Research Keyword
Field Of Study
Career
Educational Background
- 31 Mar. 2013
Keio University, Graduate School of Science and Technology, School of Integrated Design Engineering - 31 Mar. 2009
Keio University, Graduate School of Science and Technology, School of Integrated Design Engineering - 31 Mar. 2007
Keio University, Faculty of Science and Technology, Department of Electronics and Electrical Engineering
Member History
- Jun. 2024 - Jun. 2026
IEICE信号処理研究会専門委員会専門委員 - 03 Jun. 2021 - 04 Jun. 2025
IEICE英文論文誌Ａ編集委員会委員, Society - 03 Jun. 2021 - 04 Jun. 2025
IEICE和文論文誌Ａ編集委員会委員, Society - Jun. 2021 - Jun. 2024
IEICE信号処理研究会専門委員会幹事補佐, Society - Nov. 2021 - Sep. 2022
回路とシステムワークショップ実行委員会委員, Society - 15 Oct. 2021 - 31 Mar. 2022
IEICE学術奨励賞選定委員会委員, Society
Research Activity Information
Paper
- ウェーブレットベース残差拡散モデルによる単一画像超解像
朱博達; 鈴木大三; 吉田太一
IEICE 総合大会, 2025, D-11-03, Mar. 2025
Symposium, Japanese - 人物画像における逆光画像化と深層学習を用いた逆光画像強調
山本龍太; 吉田太一
Last, IEICE 信号処理シンポジウム, 2024, P1-01, Dec. 2024
Symposium, Japanese - 自己教師あり学習と画質向上事前情報を用いたノイズ除去
張晨宇; 吉田太一
Last, IEICE 総合大会, 2024, D-11A-12, Mar. 2024
Symposium, Japanese - Radiance形式HDR画像の高速二層可逆符号化
鈴木大三; 行方紗羅; 楊凱; 吉田太一
IEICE 信号処理シンポジウム, 38th, P1-05, Nov. 2023
Symposium, Japanese - 深層学習を用いた夜間撮影におけるゴースト除去
伊藤大地; 吉田太一
Last, IEICE 信号処理シンポジウム, 38th, P2-03, Nov. 2023
Symposium, Japanese - 画像反射除去のための飽和反射検出の検討,” IEICE 信号処理研究会
鈴木優奈; 吉田太一
Corresponding, IEICE 信号処理研究会, SIP1-3, Aug. 2023
Symposium, Japanese - Lightning-fast dual-layer lossless coding for radiance format high dynamic range images
Taizo Suzuki; Sara Yukikata; Kai Yang; Taichi Yoshida
IEEE Signal Process. Lett., 30, 1362-1366, 2023, Peer-reviwed
Scientific journal, English - Saturated reflection detection for reflection removal based on convolutional neural network
Taichi Yoshida; Isana Funahashi; Naoki Yamashita; Masaaki Ikehara
Lead, IEEE Access, 10, 39800-39809, Apr. 2022, Peer-reviwed
Scientific journal, English - Image adjustment for multi-exposure images based on convolutional neural networks
Isana Funahashi; Taichi Yoshida; Xi Zhang; Masahiro Iwahashi
Corresponding, IEICE Trans. Inf. & Syst., Institute of Electronics, Information and Communications Engineers (IEICE), E105-D, 1, 123-133, Jan. 2022, Peer-reviwed
Scientific journal, English - High reflection removal using CNN with detection and estimation
Isana Funahashi; Naoki Yamashita; Taichi Yoshida; Masaaki Ikehara
Corresponding, APSIPA Annual Summit Conf. 2021, 1381-1385, Dec. 2021, Peer-reviwed
International conference proceedings, English - 強反射領域の検出を用いた反射除去手法
山下尚樹; 船橋勇那; 吉田太一
Last, IEICE 総合大会, 2021, A-8-12, Mar. 2021
Symposium, Japanese - Two-layer lossless coding of HDR images specialized for radiance format
Kai Yang; Taizo Suzuki; Taichi Yoshida
APSIPA Annual Summit Conf. 2020, 1118-1121, Dec. 2020, Peer-reviwed
International conference proceedings, English - Salient object detection with importance degree
Yo Umeki; Isana Funahashi; Taichi Yoshida; Masahiro Iwahashi
Corresponding, IEEE Access, 8, 147059-147069, Aug. 2020, Peer-reviwed, © 2013 IEEE. In this article, we introduce salient object detection with importance degree (SOD-ID), which is a generalized technique for salient object detection (SOD), and propose an SOD-ID method. We define SOD-ID as a technique that detects salient objects and estimates their importance degree values. Hence, it is more effective for some image applications than SOD, which is shown via examples. The definition, evaluation procedure, and data collection for SOD-ID are introduced and discussed, and we propose its evaluation metric and data preparation, whose validity is discussed with the simulation results. Moreover, we propose an SOD-ID method, which consists of three technical blocks: instance segmentation, saliency detection, and importance degree estimation. The saliency detection block is proposed based on a convolutional neural network using the results of the instance segmentation block. The importance degree estimation block is achieved using the results of the other blocks. The proposed method accurately suppresses inaccurate saliencies and estimates the importance degree for multi-object images. In the simulations, the proposed method outperformed state-of-the-art methods with respect to the F-measure for SOD; and Spearman's and Kendall rank correlation coefficients, and the proposed metric for SOD-ID.
Scientific journal, English - Noise bias compensation for reconstruction HDR images in two layer coding
Sayaka Minewaki; Fairoza Amira Binti Hamzah; Taichi Yoshida; Masahiro Iwahashi; Hitoshi Kiya
IEEJ Trans. Electronics, Information Syst., 139, 12, 1442-1451, Dec. 2019, Peer-reviwed
Scientific journal, English - Radiance形式に基づく二層高ダイナミックレンジ画像符号化
楊凱; 鈴木大三; 吉田太一
IEICE 信号処理シンポジウム, 34th, A2-3, Nov. 2019
Symposium, Japanese - Near lossless coding of sparse histogram images based on zero-skip quantization
Sayaka Minewaki; Masahiro Iwahashi; Hiroyuki Kobayashi; Taichi Yoshida; Hitoshi Kiya
Multimed. Tools Appl., Springer New York LLC, 78, 27-45, Jan. 2019, Peer-reviwed, This paper introduces a zero-skip quantization (ZS.Q) scheme for the near lossless coding of sparse histogram images. Increases in the range of pixel values and various tone mapping operations on those pixel values mean that the histogram bins often contain no pixels. Recently, this sparseness of the histogram was used to increase the lossless coding performance by introducing histogram packing. This approach was extended to lossy coding by combining spatial quantization and lossless coding. However, such methods do not satisfy the near lossless (NL) condition. In contrast, conventional NL coding such as the JPEG-LS standard satisfies the NL condition, but does not use the histogram sparseness. In this paper, a simple ZS.Q procedure is introduced that uses the histogram sparseness to increase coding efficiency under the NL condition. The proposed method has the following advantages: 1) It guarantees the maximum quantization error is less than some threshold
2) It can be combined with an arbitrary lossless encoder, such as lossless JPEG 2000 or lossless JPEG-LS
3) Coding errors do not accumulate under repeat encoding and decoding.
Scientific journal, English - Noise bias compensation for tone mapped noisy image using prior knowledge
Sayaka Minewaki; Taichi Yoshida; Yoshinori Takei; Masahiro Iwahashi; Hitoshi Kiya
APSIPA Trans. Signal Inf. Process., 8, E3, 1-11, Jan. 2019, Peer-reviwed, © The Authors, 2019. A large number of studies have been made on denoising of a digital noisy image. In regression filters, a convolution kernel was determined based on the spatial distance or the photometric distance. In non-local mean (NLM) filters, pixel-wise calculation of the distance was replaced with patch-wise one. Later on, NLM filters have been developed to be adaptive to the local statistics of an image with introduction of the prior knowledge in a Bayesian framework. Unlike those existing approaches, we introduce the prior knowledge, not on the local patch in NLM filters but, on the noise bias (NB) which has not been utilized so far. Although the mean of noise is assumed to be zero before tone mapping (TM), it becomes non-zero value after TM due to the non-linearity of TM. Utilizing this fact, we propose a new denoising method for a tone mapped noisy image. In this method, pixels in the noisy image are classified into several subsets according to the observed pixel value, and the pixel values in each subset are compensated based on the prior knowledge so that NB of the subset becomes close to zero. As a result of experiments, effectiveness of the proposed method is confirmed.
Scientific journal, English - Safety-level estimation of aerial images based on convolutional neural network for emergency landing of unmanned aerial vehicle
Isana Funahashi; Yo Umeki; Taichi Yoshida; Masahiro Iwahashi
Corresponding, APSIPA Annual Summit Conf., 2018, 886-890, Dec. 2018, Peer-reviwed
International conference proceedings, English - ゴーストのない多露光画像合成のための深層畳み込みニューラルネットワークを用いた画像補正
船橋勇那; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 33th, C1-1, Nov. 2018
Symposium, Japanese - 画像判別を用いた顕著性物体検出
梅木陽; 瀧本裕也; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 33th, C3-4, Nov. 2018
Symposium, Japanese - 全畳み込みネットワークおよび文字列分類を用いた適応的多方向テキスト検出
千葉駿太; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 33th, C3-2, Nov. 2018
Symposium, Japanese - Co-propagation with distributed seeds for salient object detection
Yo Umeki; Taichi Yoshida; Masahiro Iwahashi
Corresponding, IEICE Trans. Inf. & Syst., Institute of Electronics, Information and Communication, Engineers, IEICE, E101-D, 6, 1640-1647, Jun. 2018, Peer-reviwed, In this paper, we propose a method of salient object detection based on distributed seeds and a co-propagation of seed information. Salient object detection is a technique which estimates important objects for human by calculating saliency values of pixels. Previous salient object detection methods often produce incorrect saliency values near salient objects in the case of images which have some objects, called the leakage of saliencies. Therefore, a method based on a co-propagation, the scale invariant feature transform, the high dimensional color transform, and machine learning is proposed to reduce the leakage. Firstly, the proposed method estimates regions clearly located in salient objects and the background, which are called as seeds and resultant seeds, are distributed over images. Next, the saliency information of seeds is simultaneously propagated, which is then referred as a co-propagation. The proposed method can reduce the leakage caused because of the above methods when the co-propagation of each information collide with each other near the boundary. Experiments show that the proposed method significantly outperforms the state-of-the-art methods in mean absolute error and F-measure, which perceptually reduces the leakage.
Scientific journal, English - Reduction of rounding noise and lifting steps in non-separable four-dimensional quadruple lifting integer wavelet transform
Fairoza Amira Binti Hamzah; Sayaka Minewaki; Taichi Yoshida; Masahiro Iwahashi
Corresponding, EURASIP J. Image Video Process., EURASIP, 2018:36, 1, 1-18, May 2018, Peer-reviwed
Scientific journal, English - Two-layer lossless coding for high dynamic range images based on range compression and adaptive inverse tone-mapping
Taichi Yoshida; Masahiro Iwahashi; Hitoshi Kiya
Lead, IEICE Trans. Fundamentals., Institute of Electronics, Information and Communication, Engineers, IEICE, E101-A, 1, 259-266, Jan. 2018, Peer-reviwed, In this paper, we propose a 2-layer lossless coding method for high dynamic range (HDR) images based on range compression and adaptive inverse tone-mapping. Recently, HDR images, which have a wider range of luminance than conventional low dynamic range (LDR) ones, have been frequently used in various fields. Since commonly used devices cannot yet display HDR images, 2-layer coding methods that decode not only HDR images but also their LDR versions have been proposed. We have previously proposed a state-of-the-art 2-layer lossless coding method for HDR images that unfortunately has huge HDR file size. Hence, we introduce two ideas to reduce the HDR file size to less than that of the previous method. The proposed method achieves high compression ratio and experiments show that it outperforms the previous method and other conventional methods.
Scientific journal, English - Pseudo multi-exposure fusion using a single image.
Yuma Kinoshita; Taichi Yoshida; Sayaka Shiota; Hitoshi Kiya
APSIPA Annual Summit Conf., 2017, 263-269, Dec. 2017, Peer-reviwed
International conference proceedings, English - Lower complexity lifting structures for hierarchical lapped transforms highly compatible with JPEG XR standard
Taizo Suzuki; Taichi Yoshida
IEEE Trans. Ciruits Syst. Video Technol., IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 27, 12, 2652-2660, Dec. 2017, Peer-reviwed, This paper presents lifting structures for hierarchical lapped transforms (HLTs) that are highly compatible with the Joint Photographic Experts Group (JPEG) eXtended Range (XR) standard and that are lower in complexity in terms of the number of operations and lifting steps than the existing HLT in JPEG XR. Two structures (T-RR and T-HH), called Householder-lifting structures, are obtained using a lifting factorization followed by a Householder factorization of a nonseparable 2D transform of rotation matrices. The third structure (T-HR) is simply derived from a combination of a Hadamard transform (T-HH) and two rotation matrices. The floating-point lifting coefficients are approximated as dyadic values, as in the existing structures of JPEG XR, because doing so costs less, thanks to the structures having only adders and shifters without multipliers. Although the new T-HH does not outperform the existing structure, the new T-RR has one fewer adder, one fewer shifter, and four fewer lifting steps than the existing one. Moreover, the new T-HR not only has one fewer adder, three fewer shifters, and two fewer lifting steps, but also can reuse T-HH, i.e., it can be used to make a more stylish codec. We show that these lower complexity HLTs are comparable in performance to the existing HLT at lossy-to-lossless image coding and at the same time highly compatible with JPEG XR.
Scientific journal, English - Four-dimensional image compression with region of interest based on non-separable double lifting integer wavelet transform
Fairoza Amira Binti Hamzah; Taichi Yoshida; Masahiro Iwahashi
Corresponding, APSIPA Annual Summit Conf., 2017, 1819-1823, Dec. 2017, Peer-reviwed
International conference proceedings, English - 複数回検出による顕著性検出
梅木陽; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 32th, C5-3, Nov. 2017
Symposium, Japanese - 完全畳み込みネットワークおよび適応的多段階クラスタリングを用いた多方向テキスト検出
千葉駿太; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 32th, C4-4, Nov. 2017
Symposium, Japanese - 多露光画像合成のための畳み込みニューラルネットワークによる画像補正
船橋勇那; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 32th, C4-1, Nov. 2017
Symposium, Japanese - Non-separable four-dimensional integer wavelet transform with reduced rounding noise
Fairoza Amira Binti Hamzah; Taichi Yoshida; Masahiro Iwahashi
Corresponding, Signal Process. Image Commun., ELSEVIER SCIENCE BV, 58, 123-133, Oct. 2017, Peer-reviwed, Since few decades ago, Discrete Cosine Transform (DCT) based digital image signal compression had been adopted as the JPEG international standard. Later, Wavelet Transform (WT) has replaced the DCT and its being applied in medical image compression. JPEG 2000, the international standardization of WT is using separable lifting structure where the multidimensional image signal is transformed separately in its horizontal and vertical direction. Besides that, each process is realized by cascading in lifting calculation. However, the necessity of waiting for previous step before calculating to the next step will make the overall delay time become longer. The delay time between input and output of WT is reduced as the proposed method reduces its lifting steps. Since the lifting step contains a rounding operation, variance of the rounding noise generated due to the rounding operation inside the transform is reduced. In this paper, unlike the conventional separable structure, the proposed non-separable structure reduces the rounding noise inside the transform, which will lead to the increasing, of coding performance. The proposed wavelet transform has a merit that its output signal, apart from the rounding noise, is exactly the same as the conventional separable structure which is a cascade of 1D structure. As a result of experiments, it was observed that the proposed method reduces the rounding noise as well as increases the performance of data compression of various 4D input signals. (C) 2017 Elsevier B.V. All rights reserved.
Scientific journal, English - 色グループ制約を用いた画像平滑化 (メディア工学)
三上陸; 吉田太一; 岩橋政宏
IEICE 画像工学研究会, 映像情報メディア学会, 87-90, Sep. 2017
Symposium, Japanese - 単一画像に基づく擬似的多重露出画像合成法
木下裕磨; 吉田太一; 塩田さやか; 貴家仁志
IEICE 画像工学研究会, 7-12, Jun. 2017
Symposium, Japanese - ストリークカメラと画像圧縮法を用いた短時間現象撮影の検討
小宮山拓磨; 吉田太一; 佐々木徹; 岩橋政宏
Corresponding, IEE プラズマ研究会, 45-49, May 2017
Symposium, Japanese - Non-separable quadruple lifting structure for four-dimensional integer wavelet transform with reduced rounding noise
Fairoza Amira Binti Hamzah; Taichi Yoshida; Masahiro Iwahashi
Corresponding, IEEE Int. Conf. Acoustics Speech Signal Process., 2017, 1148-1152, Mar. 2017, Peer-reviwed
International conference proceedings, English - High dynamic range imaging based on camera response estimation and multi-gradients fusion
Daiki Kimura; Akira Kondo; Taichi Yoshida; Masahiro Iwahashi
Corresponding, APSIPA Annual Summit Conf., 2016, 1-4, Dec. 2016, Peer-reviwed
International conference proceedings, English - Estimation method of initial labels for propagation-based saliency detection
You Umeki; Taichi Yoshida; Masahiro Iwahashi
Corresponding, APSIPA Annual Summit Conf., 2016, 1-4, Dec. 2016, Peer-reviwed
International conference proceedings, English - 事例学習辞書とスパース表現を用いたノイズ低減の多露光画像合成
坂井孝生; 井川大輔; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 31th, A3-2, Nov. 2016
Symposium, Japanese - JPEG XRのためのより効率的なリフティング構造
鈴木大三; 吉田太一
IEICE 信号処理シンポジウム, 31th, A3-1, Nov. 2016
Symposium, Japanese - 競合型ラベル伝播を基にした顕著性検出
梅木陽; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 31th, A1-1, Nov. 2016
Symposium, Japanese - Word length allocation for multiplier coefficients of minimum lifting non-separable 2D wavelet transform
Teerapong Orachon; Taichi Yoshida; Somchart Chokcha; Masahiro Iwahashi; Hitoshi Kiya
IEEJ Trans. Electronics Inf. Syst., Institute of Electrical Engineers of Japan, 136, 9, 1392-1399, Sep. 2016, Peer-reviwed, Ever since the international standard JPEG 2000 based on the lifting wavelet transform was adopted as a core technology of the digital cinema applications, its implementation issues have been discussed from various respects such as memory band width, low latency, high throughput, parallel processing and so on. Unlike the separable two-dimensional (2D) structure of the conventional lifting wavelet transform, the non-separable 2D structure can minimize the total number of lifting steps. It reduces latency of the transform on a parallel signal processing platform at the cost of introducing 2D memory accessing. However, when the word length of coefficient values inside the transform are shortened, the non-separable 2D structure has a problem that its output signal is heavily damaged by the truncation of the coefficient values. Especially, a few coefficients are extremely sensitive to the truncation in this structure. Therefore a careful treatment is required to these sensitive coefficients. In this paper, after investigating the sensitivity of each of the coefficients, a tolerable truncation error is optimally allocated to each coefficient based on the measured sensitivity. As a result, a cost function of the total word length of the wavelet transform is reduced maintaining the total amount of the noise due to truncating the coefficient values. The method is applied to expressing the coefficient values in the sum-of-power-of-two (SPT) format for a low computational load implementation.
Scientific journal, English - Channel scaling for integer implementation of minimum lifting 2D wavelet transform
Teerapong Orachon; Taichi Yoshida; Somchart Chokchaitam; Masahiro Iwahashi; Hitoshi Kiya
IEICE Trans. Fundamentals., IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E99-A, 7, 1420-1429, Jul. 2016, Peer-reviwed, The lifting wavelet transform (WT) has been widely applied to image coding. Recently, the total number of lifting steps has been minimized introducing a non-separable 2D structure so that delay from input to output can be reduced in parallel processing. However the minimum lifting WT has a problem that its upper bound of the rate-distortion curve is lower than that of the standard lifting WT. This is due to the rounding noise generated inside the transform in its integer implementation. This paper reduces the rounding noise introducing channel scaling. The channel scaling is designed so that the dynamic range of signal values is fully utilized at each channel inside the transform. As a result, the signal to noise ratio is increased and therefore the upper bound of the minimum lifting WT in lossy coding is improved.
Scientific journal, English - Subset noise bias compensation for tone-mapping and up-scaling of JPEG images
Masahiro Iwahashi; Tin Moe Aye; Taichi Yoshida; Hitoshi Kiya
IEEE Int. Symp. Circuits Syst., 2016, 2226-2229, May 2016, Peer-reviwed
International conference proceedings, English - Optimum SPT allocation for multipliers of minimum lifting 2D wavelet transform
Masahiro Iwahashi; Teerapong Orachon; Tin Moe Aye; Taichi Yoshida; Hitoshi Kiya
IEEE Int. Symp. Circuits Syst., 2016, 2543-2546, May 2016, Peer-reviwed
International conference proceedings, English - Adaptive directional lifting structure of three dimensional non-separable discrete wavelet transform for high resolution volumetric data compression
Fairoza Amira Binti Hamzah; Taichi Yoshida; Masahiro Iwahashi; Hitoshi Kiya
Corresponding, IEICE Trans. Fundamentals., IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E99-A, 5, 892-899, May 2016, Peer-reviwed, As three dimensional (3D) discrete wavelet transform (DWT) is widely used for high resolution volumetric data compression, and to further improve the performance of lossless coding, the adaptive directional lifting (ADL) structure based on non-separable 3D DWT with a (5,3) filter is proposed in this paper. The proposed 3D DWT has less lifting steps and better prediction performance compared to the existing separable 3D DWT with fixed filter coefficients. It also has compatibility with the conventional DWT defined by the JPEG2000 international standard. The proposed method shows comparable and better results with the non separable 3D DWT and separable 3D DWT and it is effective for lossless coding of high resolution volumetric data.
Scientific journal, English - Single image super resolution by L2 approximation with random sampled dictionary
Takanori Fujisawa; Taichi Yoshida; Kazu Mishiba; Masaaki Ikehara
IEICE Trans. Fundamentals., IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E99-A, 2, 612-620, Feb. 2016, Peer-reviwed, In this paper, we propose an example-based single image super resolution (SR) method by l(2) approximation with self-sampled image patches. Example-based super resolution methods can reconstruct high resolution image patches by a linear combination of atoms in an overcomplete dictionary. This reconstruction requires a pair of two dictionaries created by tremendous low and high resolution image pairs from the prepared image databases. In our method, we introduce the dictionary by random sampling patches from just an input image and eliminate its training process. This dictionary exploits the self-similarity of images and it will no more depend on external image sets, which consern the storage space or the accuracy of referred image sets. In addition, we modified the approximation of input image to an l(2)-norm minimization problem, instead of commonly used sparse approximation such as l(1)-norm regularization. The l(2) approximation has an advantage of computational cost by only solving an inverse problem. Through some experiments, the proposed method drastically reduces the computational time for the SR, and it provides a comparable performance to the conventional example-based SR methods with an l(1) approximation and dictionary training.
Scientific journal, English - Noise bias compensation based on bayesian inference for tone mapped noisy image
Masahiro Iwahashi; Fairoza Amira Binti Hamzah; Taichi Yoshida; Hitoshi Kiya
APSIPA Annual Summit Conf., 2015, 440-443, Dec. 2015, Peer-reviwed
International conference proceedings, English - Two layer coding of HDR images with noise bias compensation
Masahiro Iwahashi; Fairoza Amira Binti Hamzah; Taichi Yoshida; Hitoshi Kiya
APSIPA Annual Summit Conf., 2015, 436-439, Dec. 2015, Peer-reviwed
International conference proceedings, English - Channel scaling for rounding noise reduction in minimum lifting 3D wavelet transfrom
Fairoza Amira Binti Hamzah; Taichi Yoshida; Masahiro Iwahashi; Hitoshi Kiya
Corresponding, APSIPA Annual Summit Conf., 2015, 888-891, Dec. 2015, Peer-reviwed
International conference proceedings, English - レスポンスカーブ関数推定と合成勾配正則化による高ダイナミックレンジ画像生成
木村大輝; 吉田太一; 岩橋政宏
Corresponding, IEICE 信号処理シンポジウム, 30th, A4-3, Nov. 2015
Symposium, Japanese - Hybrid method for multi-exposure image fusion based on weighted mean and sparse representation
Takao Sakai; Daiki Kimura; Taichi Yoshida; Masahiro Iwahashi
Corresponding, European Signal Process. Conf., 2015, 814-818, Sep. 2015, Peer-reviwed
International conference proceedings, English - Bit-depth scalable lossless coding for high dynamic range images
Masahiro Iwahashi; Taichi Yoshida; Norrima Binti Mokhtar; Hitoshi Kiya
EURASIP J. Adv. Signal Process., SPRINGER INTERNATIONAL PUBLISHING AG, 2015, 22, 1-15, Mar. 2015, Peer-reviwed, In this paper, we propose a bit-depth scalable lossless coding method for high dynamic range (HDR) images based on a reversible logarithmic mapping. HDR images are generally expressed as floating-point data, such as in the OpenEXR or RGBE formats. Our bit-depth scalable coding approach outputs base layer data and enhancement layer data. It can reconstruct the low dynamic range (LDR) image from the base layer data and reconstructs the HDR image by adding the enhancement layer data. Most previous two-layer methods have focused on the lossy coding of HDR images. Unfortunately, the extension of previous lossy methods to lossless coding does not significantly compress the enhancement layer data. This is because the bit depth becomes very large, especially for HDR images in floating-point data format. To tackle this problem, we apply a reversible logarithmic mapping to the input HDR data. Moreover, we introduce a format conversion to avoid any degradation in the quality of the reconstructed LDR image. The proposed method is effective for both OpenEXR and RGBE formats. Through a series of experiments, we confirm that the proposed method decreases the volume of compressed data while maintaining the visual quality of the reconstructed LDR images.
Scientific journal, English - L2 norm optimization of tone mapping for two layer lossless coding of HDR images
Masahiro Iwahashi; Taichi Yoshida; Hitoshi Kiya
APSIPA Annual Summit Conf., 2014, 1-4, Dec. 2014, Peer-reviwed
International conference proceedings, English - Non-separable 3D integer wavelet transform for lossless data compression
Teerapong Orachon, Suvit Poomrittigul, Taichi Yoshida, Masahiro Iwahashi, and Somchart Chokchaitam
Corresponding, Science J. Circuits Syst. Signal Process., 3, 6, 35-46, Dec. 2014, Peer-reviwed, False
Scientific journal, English - Range reduction of HDR images for backward compatibility with LDR image processing
Masahiro Iwahashi; Taichi Yoshida; Hitoshi Kiya
APSIPA Annual Summit Conf., 2014, 1-4, Dec. 2014, Peer-reviwed
International conference proceedings, English - Non-separable three dimensional discrete wavelet transform with adaptive directional prediction
Fairoza Amira Binti Hamzah; Teerapong Orachon; Taichi Yoshida; Masahiro Iwahashi; Hitoshi Kiya
Corresponding, APSIPA Annual Summit Conf., 2014, 1-4, Dec. 2014, Peer-reviwed
International conference proceedings, English - Image super-resolution based on shock filter and non-local means
Taichi Yoshida; Masaaki Ikehara
Lead, APSIPA Annual Summit Conf., 2014, 1-4, Dec. 2014, Peer-reviwed
International conference proceedings, English - 画像平均とTVを用いた逆畳み込みによる画像のノイズ除去
豊田真帆; 池原雅章; 吉田太一
IEICE 信号処理シンポジウム, 29th, P2-8, Nov. 2014
Symposium, Japanese - ヒストグラムパッキングとグラデーション予測を用いた二階層ロスレスHDR画像符号化
長谷川治郎; 吉田太一; 岩橋政宏; 貴家仁志
Corresponding, IEICE 信号処理シンポジウム, 29th, P2-4, Nov. 2014
Symposium, Japanese - Super-resolutions based on shock filter and non local means with BM3D for congruity
Ken Iwamoto; Taichi Yoshida; Masaaki Ikehara
Corresponding, TENCON - IEEE Region 10 Conf., 2014, 1-6, Oct. 2014, Peer-reviwed
International conference proceedings, English - Adaptive reversible data hiding in frequency domain via integer-to-integer transform
Taichi Yoshida; Taizo Suzuki; Masaaki Ikehara
Lead, IEICE Trans. Fundamentals., IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E97-A, 1, 384-392, Jan. 2014, Peer-reviwed, We propose an adaptive reversible data hiding method with superior visual quality and capacity in which an adaptive generalized difference expansion (AGDE) method is applied to an integer-to-integer subband transform (I2I-ST). I2I-ST performs the reversible subband transform and the AGDE method is a state-of-the-art method of reversible data hiding. The results of experiments we performed objectively and perceptually show that the proposed method has better visual quality than conventional methods at the same embedding rate due to low variance in the frequency domain.
Scientific journal, English - 自己相似性と自己合同性を用いた画像の超解像
岩本健; 吉田太一; 池原雅章
Corresponding, IEICE 信号処理シンポジウム, 28th, B4-4, Nov. 2013
Symposium, Japanese - 1次元方向性冗長フィルタバンクと画像のスパース性を用いた合成による画像ノイズ除去
吉田太一; 古橋直樹; 池原雅章
Lead, IEICE 信号処理シンポジウム, 28th, B1-1, Nov. 2013
Symposium, Japanese - Colorization based on piecewise autoregressive model
Yasuhiro Nakajima; Takashi Ueno; Taichi Yoshida; Masaaki Ikehara
Asilomar Conf. Signals Syst. Comput., 2013, 1990-1994, Nov. 2013, Peer-reviwed
International conference proceedings, English - Image super-resolution method based on non-local means and self similarity
Taichi Yoshida; Tomoya Murakami; Masaaki Ikehara
Lead, IEEE Int. Symp. Intell. Signal Process. Commun. Syst., 2013, 509-512, Nov. 2013, Peer-reviwed
International conference proceedings, English - Image denoising by adaptive directional lifting-based discrete wavelet transform and quantization
Naoki Furuhashi; Azusa Oota; Taichi Yoshida; Masaaki Ikehara
Asilomar Conf. Signals Syst. Comput., 2013, 1995-1999, Nov. 2013, Peer-reviwed
International conference proceedings, English - Blocking artifacts reduction of DCT compressed image based on block wiener filtering
Azusa Ota; Taichi Yoshida; Masaaki Ikehara
Corresponding, IEEE Int. Conf. Adv. Technol. Commun., 2013, 85-89, Oct. 2013, Peer-reviwed
International conference proceedings, English - Two dimensional M-channel non-separable filter banks based on cosine modulated filter banks with diagonal shifts
Taichi Yoshida; Seisuke Kyochi; Masaaki Ikehara
Lead, IEICE Trans. Fundamentals., IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E96-A, 8, 1685-1694, Aug. 2013, Peer-reviwed, In this paper, we propose a new class of two dimensional (2D) M-channel (M-ch) non-separable filter banks (FBs) based on cosine modulated filter banks (CMFBs) via a new diagonally modulation scheme. Until now, many researchers have proposed 2D non-separable CMFBs. Nevertheless, efficient direction-selective CMFBs have not been yet. Thanks to our new modulations with diagonal shifts, proposed CMFBs have several frequency supports including direction-selective ones which cannot be realized by conventional ones. In a simulation, we show design examples of proposed CMFBs and their various directional frequency supports.
Scientific journal, English - Rate-dependent seam carving and its application to content-aware image coding
Yuichi Tanaka; Taichi Yoshida; Madoka Hasegawa; Shigeo Kato; Masaaki Ikehara
APSIPA Trans. Signal Inf. Process., Cambridge University Press, 2, 1-15, Feb. 2013, Peer-reviwed, Content-aware image resizing (CAIR) is desired because it preserves prominent regions in a resized image. However, CAIR requires high computational complexity to perform in mobile devices, though it is desired for these displays. Moreover, transmitting the side information for CAIR from the encoder is a problem since it usually requires high bitrates compared with those for image data. In this paper, we present a rate-dependent CAIR method that produce various retargeting results based on the bitrates for side information. Furthermore, we apply the proposed technique to wavelet-based image coding. Our proposed content-aware image coding method provides good performances for both CAIR and image coding. © The Authors, 2013.
Scientific journal, English - Adaptive reversible data hiding in frequency domain via integer-to-integer transform
Taichi Yoshida; Yusuke Okamura; Taizo Suzuki; Masaaki Ikehara
Lead, APSIPA Annual Summit Conf., 2012, 1-4, Dec. 2012, Peer-reviwed
International conference proceedings, English - Color image coding based on the colorization
Takashi Ueno; Taichi Yoshida; Masaaki Ikehara
Corresponding, APSIPA Annual Summit Conf., 2012, 1-4, Dec. 2012, Peer-reviwed
International conference proceedings, English - Wavelet-based content-aware image coding with rate-dependent seam carving
Taichi Yoshida; Yuichi Tanaka; Madoka Hasegawa; Shigeo Kato; Masaaki Ikehara
Lead, IEEE Int. Conf. Image Process., 2012, 2509-2512, Sep. 2012, Peer-reviwed
International conference proceedings, English - Two dimensional non-separable adaptive directional lifting structure of discrete wavelet transform
Taichi Yoshida; Taizo Suzuki; Seisuke Kyochi; Masaaki Ikehara
Lead, IEICE Trans. Fundamentals., IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E94-A, 10, 1920-1927, Oct. 2011, Peer-reviwed, In this paper, we propose a two dimensional (2D) nonseparable adaptive directional lifting (ADL) structure for discrete wavelet transform (DWT) and its image coding application. Although a 2D nonseparable lifting structure of 9/7 DWT has been proposed by interchanging some lifting, we generalize a polyphase representation of 2D non-separable lifting structure of DWT. Furthermore, by introducing the adaptive directional filteringingto the generalized structure, the 2D non-separable ADL structure is realized and applied into image coding. Our proposed method is simpler than the ID ADL, and can select the different transforming direction with ID ADL. Through the simulations, the proposed method is shown to be efficient for the lossy and lossless image coding performance.
Scientific journal, English - Two dimensional non-separable adaptive directional lifting structure of discrete wavelet transform
Taichi Yoshida; Taizo Suzuki; Seisuke Kyochi; Masaaki ikehara
Lead, IEEE Int. Conf. Acoustics Speech Signal Process., 2011, 1529-1532, May 2011, Peer-reviwed
International conference proceedings, English - A simplified lattice structure of two dimensional generalized lapped orthogonal transform
Taichi Yoshida; Seisuke Kyochi; Masaaki Ikehara
Lead, IEICE Trans. Fundamentals., IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E94-A, 2, 671-679, Feb. 2011, Peer-reviwed, In this paper, we propose a novel lattice structure of two dimensional (2D) nonseparable linear-phase paraunitary filter banks (LPPUFBs) called 2D GenLOT. Muramatsu et al. have previously proposed a lattice structure of 2D nonseparable LPPUFBs which have efficient frequency response. However, the proposed structure requires less number of design parameters and computational costs than the conventional one. Through some design examples and simulation results, we show that both filter banks have comparable frequency response and coding gain.
Scientific journal, English - 2次元非可分型離散ウェーブレット変換の方向適応型リフティング実現
吉田太一; 鈴木大三; 京地清介; 池原雅章
Lead, IEICE 信号処理シンポジウム, 25th, C1-2, Nov. 2010
Symposium, Japanese - A simplified lattice structure of two-dimensional generalized lapped orthogonal transform (2-D GenLOT) for image coding
Taichi Yoshida; Seisuke Kyochi; Masaaki Ikehara
Lead, IEEE Int. Conf. Image Process., 2010, 349-352, Sep. 2010, Peer-reviwed
International conference proceedings, English - A generalized lapped transform with unequal length filters based on first-order linear-phase filter banks
Taichi Yoshida; Yuichi Tanaka; Masaaki Ikehara
Lead, Asilomar Conf. Signals Syst. Comput., 2007, 1083-1087, Nov. 2007, Peer-reviwed
International conference proceedings, English
Courses
- 画像処理工学
Apr. 2025 - Present - マルチメディア信号処理特論
Apr. 2025 - Present - 情報通信工学実験A&電子情報学実験A
Apr. 2018 - Present
The University of Electro-Communications - 情報データ解析論
Apr. 2024 - Mar. 2025 - 基礎科学実験A
Apr. 2019 - Mar. 2025
The University of Electro-Communications - 情報・通信工学基礎
Oct. 2018 - Mar. 2019
The University of Electro-Communications - 先端工学基礎課程輪講AB
Apr. 2018 - Sep. 2018
The University of Electro-Communications
Research Themes
- 逆光画像化を用いた深層学習に基づく逆光画像強調
吉田 太一
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (C), 電気通信大学, Grant-in-Aid for Scientific Research (C), Principal investigator, 24K15007
Apr. 2024 - Mar. 2027 - Two-Layer Coding for High Dynamic Range Images Based on Histogram Sparsity
Taichi Yoshida
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Young Scientists (B), Grant-in-Aid for Young Scientists (B), Principal investigator, In this research, techniques are proposed for two-layer coding of high dynamic range (HDR) images. The two-layer coding is applied by JPEG XT that is a recent global standard of HDR image coding. The proposed techniques improve compression rates for the two-layer coding as fundamentals and are selectively used without interfering other fundamental ones. They are realized based on the particular property of HDR images and their two-layer coding that is introduced in this research. Therefore, this research contributes the next global standard for HDR image coding. In simulation, they show the improvement of compression rates, which reaches up to 15%, compared with methods without themselves., 16K18104
Apr. 2016 - Mar. 2020 - Multi-dimensional visualization of ultra-fast optical emission based on streak camera and sparse image processing
Masahiro Iwahashi; Toru Sasaki; YOSHIDA TAICHI
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Challenging Exploratory Research, Nagaoka University of Technology, Grant-in-Aid for Challenging Exploratory Research, Recently, it became possible to observe the ultra-fast optical emission phenomenon. It has been difficult to observe spatial observation of the phenomenon with conventional approaches. This project makes it possible to observe the phenomenon as a multi-dimensional information with a compact equipment based on a streak camera. We developed 1) an observation equipment with the polarization optical devices, 2) sparse coding algorithm to reconstruct the multi-dimensional information, and 3) a compact system based on a parallel computing system., 16K13715
Apr. 2016 - Mar. 2019 - Two-Layer Lossless Coding for High Dynamic Range Images
Taichi Yoshida
公益財団法人ユニオンツール, 育英奨学会, 長岡技術科学大学, Principal investigator
Apr. 2015 - Mar. 2018 - Functionally layered coding for visual odometry and its application to autonomous traveling camera
Iwahashi Masahiro; Seiji Kamimura; Tetsuya Kimura; Taichi Yoshida
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B), Nagaoka University of Technology, Grant-in-Aid for Scientific Research (B), It is required to develop a group of crawler robots which can autonomously travel and monitor around a rescue scene. In such a scene, there is no well prepared infrastructure for telecommunications for the autonomous travel robots. This research develops a new video processing method for visual odometry based on high dynamic range image processing techniques. It also develops a new data compression method for a narrow band telecommunication network based on the functionally layered coding technique., 26289117
Apr. 2014 - Mar. 2017 - Example-based hierarchical dictionary learning for image super-resolution based on image self-similarity
Taichi Yoshida
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Research Activity Start-up, Nagaoka University of Technology, Grant-in-Aid for Research Activity Start-up, This research is focused on the construction of an example-based hierarchical dictionary and its application for image super-resolution based on image self-similarity. The proposed dictionary constructs a hierarchical dictionary and relates a degraded atom with a multiple fine atom to represent various images from one degraded image. Based on the image self-similarity, we search similar neighbors and approximate target and neighbor patches with the dictionary and same coefficients under weights calculated from similarities of them. The proposed method save calculation costs and produces fine resultant images., 26889031
Aug. 2014 - Mar. 2016 - 2次元方向性フィルタバンクの実現と画像の変換符号化への応用
吉田 太一
日本学術振興会, 科学研究費助成事業, 慶應義塾大学, 特別研究員奨励費, 本年度は研究実施計画に基づき、昨年度に実現した方向を考慮した周波数変換を実現する2次元方向性フィルタバンクの周波数変調構造を基にした設計法の多分割への一般化を確立し、画像符号化およびノイズ除去への応用を実現した。 従来の設計法による2次元方向性フィルタバンクは、多様な周波数分割形状を実現しており良好な変換性能を達成しているが、構造上フィルタサイズを小さくできないために、変換の局所性が低く且つ変換に要する加算器や乗算器数が膨大であり非実用的である。そこで、低演算な方向性フィルタバンクを実現するために、周波数変調構造を基にした設計法を提案する。周波数変調構造は、1つのプロトタイプフィルタとそれを分割数分変調させる変調行列により構成され、変調行列の演算量は低く、フィルタサイズおよび変換に要する演算量はプロトタイプフィルタに依存し、1つのフィルタによる変換と同等にまで演算量を削減できる。 申請者はまず、昨年度実現した斜め方向への変調方式による設計法を多分割へ拡張した。提案法は、変調行列に周波数位相特性にのみ影響を与える項を加え、位相変調により不正成分が互いに打ち消しあう条件を導出した。これにより、入力信号と再構成信号の誤差を構造的に無くし、多分割の一般化された設計法を確立した。また画像符号化およびノイズ除去に応用し、従来法と比較してほぼ同等の画質を実現しつつ、加算器および乗算器数を概算で約80パーセント削減した。, 12J07215
2012 - 2013
