Convolutional Dictionary Pair Learning Network for Image Representation Learning.pdf

Zhang, Zhao; Sun, Yulin; Wang, Yang; Zha, Zheng-Jun; Yan, Shuicheng; Wang, Meng

doi:10.36227/techrxiv.11459574.v1

Cited by 3 publications

(3 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The training datasets include BSD400 [3] and WED [38] for denoising by adding Gaussian noise levels σ ∈ {15, 25, 50}, Rain100L [64] for deraining, and SOTS [28] for dehazing. In addition, we include two challenging tasks of image deblurring and low-light enhancement and use the GoPro [41] and LOL dataset [59] for training, as previous research [8,69,70]. For all the datasets, we follow the standard practices in data splitting and preprocessing in the field.…”

Section: Methodsmentioning

confidence: 99%

Transfer Learning in Motor Imagery Brain Computer Interface: A Review

2022

J. Shanghai Jiaotong Univ. (Sci.)

View full text Add to dashboard Cite

Sparse-view computed tomography (CT)-using a small number of projections for tomographic reconstructionenables much lower radiation dose to patients and accelerated data acquisition. The reconstructed images, however, suffer from strong artifacts, greatly limiting their diagnostic value. Current trends for sparse-view CT turn to the raw data for better information recovery. The resultant dual-domain methods, nonetheless, suffer from secondary artifacts, especially in ultra-sparse view scenarios, and their generalization to other scanners/protocols is greatly limited. A crucial question arises: have the image post-processing methods reached the limit? Our answer is not yet. In this paper, we stick to image post-processing methods due to great flexibility and propose global representation(GloRe) distillation framework for sparse-view CT, termed GloReDi. First, we propose to learn GloRe with Fourier convolution, so each element in GloRe has an image-wide receptive field. Second, unlike methods that only use the full-view images for supervision, we propose to distill GloRe from intermediate-view reconstructed images that are readily available but not explored in previous literature. The success of GloRe distillation is attributed to two key components: representation directional distillation to align the GloRe directions, and band-pass-specific contrastive distillation to gain clinically important details. Extensive experiments demonstrate the superiority of the proposed GloReDi over the state-of-the-art methods, including dual-domain ones. The source code is available at https: //github.com/longzilicart/GloReDi.

show abstract

Section: Methodsmentioning

confidence: 99%

Transfer Learning in Motor Imagery Brain Computer Interface: A Review

2022

J. Shanghai Jiaotong Univ. (Sci.)

View full text Add to dashboard Cite

show abstract

“…SDRCF [120]. SDRCF is inspired by the sparse representation (SR) [56][57][58][59][60][61][62][63][64][65][66], which simultaneously incorporates the local geometrical structures of both the data and features into CF, and obtain a weight matrix. For a sample x and a matrix DN   containing the dictionary atoms in its columns, SR represents x using as few entries of as possible, defined as follows:…”

Section: Scf [127] and Rscf [127]mentioning

confidence: 99%

“…Because the RL methods can effectively simplify the complex input data, eliminate invalid information and extract useful information (or features) from observed inputs [45][46][47][48][49][50][51][52][53][54][55]. Classical RL approaches include feature extraction (FE) , sparse dictionary learning (SDL) [56][57][58][59][60][61][62][63][64][65][66], low-rank coding (LRC) [91][92][93][94][95][96][97][98][99][100][101][102][103][104], matrix factorization (MF) [1][2][3][4][5][6] [ [105][106][107][108][109][110][111]…”

Section: Introductionmentioning

confidence: 99%