检索结果-内蒙古大学图书馆

Collaborative Multi-View Convolutions With Gating For Accurate And Fast Volumetric Medical Image Segmentation

学校读者我要写书评

暂无评论

Collaborative Multi-View Convolutions With Gating For Accura...

IEEE International Symposium on Biomedical Imaging

作者： Cheng Li Jin Ye Junjun He Shanshan Wang Lixu Gu Yu Qiao Paul C. Lauterbur Research Center for Biomedical Imaging SIAT CAS Shenzhen China Shenzhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab SIAT CAS Shenzhen China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen China School of Biomedical Engineering/the Institute of Medical Robotics Shanghai Jiao Tong University Shanghai China

ISBN: (纸本)9781665412469;9781665429474

Due to their high capacity in capturing 3D spatial information, 3D Fully Convolutional Neural Networks (3D FCNs), especially 3D U-Net, are prevalent for volumetric medical image segmentation. However, 3D convolutions are much more computationally complex than 2D convolutions and thus, are more prone to overfitting. This paper proposes Collaborative Multi-View convolutions (CMV convs) that can keep the model complexity similar to those employing 2D convolutions while capturing the 3D spatial context like 3D convolutions. Specifically, CMV convs simultaneously extract information from three orthogonal views with three parameter-shared 2D convolutions. A Global-Guided Gating mechanism (3G) is further designed that selectively passes information from CMV convs to the next stage. Combined with 3G, a CMV conv become a G-CMV conv that constitutes a plug-and-play module, which can be easily integrated into various 3D CNNs for image segmentation. Extensive experiments utilizing BraTS18 dataset have been conducted. Our method achieves competitive results compared to state-of-the-art methods with over 10× fewer parameters than 3D-UNet.

关键词： Solid modeling Image segmentation Interpolation Three-dimensional displays Computational modeling Biological system modeling Collaboration

Tensor Low-Rank Reconstruction for Semantic Segmentation

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Chen, Wanli Zhu, Xinge Sun, Ruoqi He, Junjun Li, Ruiyu Shen, Xiaoyong Yu, Bei Chinese University of Hong Kong Hong Kong Shanghai Jiao Tong University China ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China SmartMore United States

Context information plays an indispensable role in the success of semantic segmentation. Recently, non-local self-attention based methods are proved to be effective for context information collection. Since the desired context consists of spatial-wise and channel-wise attentions, 3D representation is an appropriate formulation. However, these non-local methods describe 3D context information based on a 2D similarity matrix, where space compression may lead to channel-wise attention missing. An alternative is to model the contextual information directly without compression. However, this effort confronts a fundamental difficulty, namely the high-rank property of context information. In this paper, we propose a new approach to model the 3D context representations, which not only avoids the space compression but also tackles the high-rank difficulty. Here, inspired by tensor canonical-polyadic decomposition theory (i.e, a high-rank tensor can be expressed as a combination of rank-1 tensors.), we design a low-rank-to-high-rank context reconstruction framework (i.e, RecoNet). Specifically, we first introduce the tensor generation module (TGM), which generates a number of rank-1 tensors to capture fragments of context feature. Then we use these rank-1 tensors to recover the high-rank context features through our proposed tensor reconstruction module (TRM). Extensive experiments show that our method achieves state-of-the-art on various public datasets. Additionally, our proposed method has more than 100 times less computational cost compared with conventional non-local-based methods. Copyright © 2020, The Authors. All rights reserved.

关键词： Tensors

Blind super-resolutionwith iterative kernel correction

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Gu, Jinjin Lu, Hannan Zuo, Wangmeng Dong, Chao School of Science and Engineering Chinese University of Hong Kong Shenzhen China School of Computer Science and Technology Harbin Institute of Technology Harbin China ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences

Deep learning based methods have dominated superresolution (SR) field due to their remarkable performance in terms of effectiveness and efficiency. Most of these methods assume that the blur kernel during downsampling is predefined/ known (e.g., bicubic). However, the blur kernels involved in real applications are complicated and unknown, resulting in severe performance drop for the advanced SR methods. In this paper, we propose an Iterative Kernel Correction (IKC) method for blur kernel estimation in blind SR problem, where the blur kernels are unknown. We draw the observation that kernel mismatch could bring regular artifacts (either over-sharpening or over-smoothing), which can be applied to correct inaccurate blur kernels. Thus we introduce an iterative correction scheme - IKC that achieves better results than direct kernel estimation. We further propose an effective SR network architecture using spatial feature transform (SFT) layers to handle multiple blur kernels, named SFTMD. Extensive experiments on synthetic and real-world images show that the proposed IKC method with SFTMD can provide visually favorable SR results and the state-of-the-art performance in blind SR problem. Copyright © 2019, The Authors. All rights reserved.

关键词： Network architecture

Reflash Dropout in Image Super-Resolution

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Kong, Xiangtao Liu, Xina Gu, Jinjin Qiao, Yu Dong, Chao ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab. Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China University of Chinese Academy of Sciences China The University of Sydney Australia Shanghai AI Laboratory Shanghai China

Dropout is designed to relieve the overfitting problem in high-level vision tasks but is rarely applied in low-level vision tasks, like image super-resolution (SR). As a classic regression problem, SR exhibits a different behaviour as high-level tasks and is sensitive to the dropout operation. However, in this paper, we show that appropriate usage of dropout benefits SR networks and improves the generalization ability. Specifically, dropout is better embedded at the end of the network and is significantly helpful for the multi-degradation settings. This discovery breaks our common sense and inspires us to explore its working mechanism. We further use two analysis tools – one is from a recent network interpretation work, and the other is specially designed for this task. The analysis results provide side proofs to our experimental findings and show us a new perspective to understand SR networks. Copyright © 2021, The Authors. All rights reserved.

关键词： Optical resolving power

Suppressing model overfitting for image super-resolution networks

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Feng, Ruicheng Gu, Jinjin Qiao, Yu Dong, Chao ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences School of Science and Engineering Chinese University of Hong Kong Shenzhen Hong Kong Chinese University of Hong Kong Hong Kong

Large deep networks have demonstrated competitive performance in single image super-resolution (SISR), with a huge volume of data involved. However, in real-world scenarios, due to the limited accessible training pairs, large models exhibit undesirable behaviors such as overfitting and memorization. To suppress model overfitting and further enjoy the merits of large model capacity, we thoroughly investigate generic approaches for supplying additional training data pairs. In particular, we introduce a simple learning principle MixUp [43] to train networks on interpolations of sample pairs, which encourages networks to support linear behavior in-between training samples. In addition, we propose a data synthesis method with learned degradation, enabling models to use extra high-quality images with higher content diversity. This strategy proves to be successful in reducing biases of data. By combining these components - MixUp and synthetic training data, large models can be trained without overfitting under very limited data samples and achieve satisfactory generalization performance. Our method won the second place in NTIRE2019 Real SR Challenge. Copyright © 2019, The Authors. All rights reserved.

关键词： Optical resolving power

Self-slimmed vision Transformer

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zong, Zhuofan Li, Kunchang Song, Guanglu Wang, Yali Qiao, Yu Leng, Biao Liu, Yu School of Computer Science and Engineering Beihang University China SenseTime Research China ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China University of Chinese Academy of Sciences China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society China Shanghai AI Laboratory China

vision transformers (ViTs) have become the popular structures and outperformed convolutional neural networks (CNNs) on various vision tasks. However, such powerful transformers bring a huge computation burden, because of the exhausting token-to-token comparison. The previous works focus on dropping insignificant tokens to reduce the computational cost of ViTs. But when the dropping ratio increases, this hard manner will inevitably discard the vital tokens, which limits its efficiency. To solve the issue, we propose a generic self-slimmed learning approach for vanilla ViTs, namely SiT. Specifically, we first design a novel Token Slimming Module (TSM), which can boost the inference efficiency of ViTs by dynamic token aggregation. As a general method of token hard dropping, our TSM softly integrates redundant tokens into fewer informative ones. It can dynamically zoom visual attention without cutting off discriminative token relations in the images, even with a high slimming ratio. Furthermore, we introduce a concise Feature Recalibration Distillation (FRD) framework, wherein we design a reverse version of TSM (RTSM) to recalibrate the unstructured token in a flexible auto-encoder manner. Due to the similar structure between teacher and student, our FRD can effectively leverage structure knowledge for better convergence. Finally, we conduct extensive experiments to evaluate our SiT. It demonstrates that our method can speed up ViTs by 1.7× with negligible accuracy drop, and even speed up ViTs by 3.6× while maintaining 97% of their performance. Surprisingly, by simply arming LV-ViT with our SiT, we achieve new state-of-the-art performance on ImageNet. Code is available at https://***/Sense-X/SiT. © 2021, CC BY.

关键词： Distillation

TTPP: Temporal transformer with progressive prediction for efficient action anticipation

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wang, Wen Peng, Xiaojiang Su, Yanzhou Qiao, Yu Cheng, Jian School of Information and Communication Engineering University of Electronic Science and Technology of China Chengdu Sichuan611731 China ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab. Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society

Video action anticipation aims to predict future action categories from observed frames. Current state-of-the-art approaches mainly resort to recurrent neural networks to encode history information into hidden states, and predict future actions from the hidden representations. It is well known that the recurrent pipeline is inefficient in capturing long-term information which may limit its performance in predication task. To address this problem, this paper proposes a simple yet efficient Temporal Transformer with Progressive Prediction (TTPP) framework, which repurposes a Transformer-style architecture to aggregate observed features, and then leverages a light-weight network to progressively predict future features and actions. Specifically, predicted features along with predicted probabilities are accumulated into the inputs of subsequent prediction. We evaluate our approach on three action datasets, namely TVSeries, THUMOS-14, and TV-Human-Interaction. Additionally we also conduct a comprehensive study for several popular aggregation and prediction strategies. Extensive results show that TTPP not only outperforms the state-of-the-art methods but also more efficient. Copyright © 2020, The Authors. All rights reserved.

关键词： Forecasting

SpiderCNN: Deep learning on point sets with parameterized convolutional filters

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Xu, Yifan Fan, Tianqi Xu, Mingye Zeng, Long Qiao, Yu Tsinghua University Guangdong Key Lab of Computer Vision and Virtual Reality SIAT-SenseTime Joint Lab Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences

Deep neural networks have enjoyed remarkable success for various vision tasks, however it remains challenging to apply CNNs to domains lacking a regular underlying structures such as 3D point clouds. Towards this we propose a novel convolutional architecture, termed SpiderCNN, to efficiently extract geometric features from point clouds. SpiderCNN is comprised of units called SpiderConv, which extend convolutional operations from regular grids to irregular point sets that can be embedded in Rn, by parametrizing a family of convolutional filters. We design the filter as a product of a simple step function that captures local geodesic information and a Taylor polynomial that ensures the expressiveness. SpiderCNN inherits the multi-scale hierarchical architecture from classical CNNs, which allows it to extract semantic deep features. Experiments on ModelNet40[4] demonstrate that SpiderCNN achieves state-of-the-art accuracy 92.4% on standard benchmarks, and shows competitive performance on segmentation task. Copyright © 2018, The Authors. All rights reserved.

关键词： Convolution

A new journey from SDRTV to HDRTV

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Chen, Xiangyu Zhang, Zhengwen Ren, Jimmy S. Tian, Lynhoo Qiao, Yu Dong, Chao ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shanghai China SenseTime Research Shanghai China Qing Yuan Research Institute Shanghai Jiao Tong University Shanghai China Shanghai AI Laboratory Shanghai China

Nowadays modern displays are capable to render video content with high dynamic range (HDR) and wide color gamut (WCG). However, most available resources are still in standard dynamic range (SDR). Therefore, there is an urgent demand to transform existing SDR-TV contents into their HDR-TV versions. In this paper, we conduct an analysis of SDRTV-to-HDRTV task by modeling the formation of SDRTV/HDRTV content. Base on the analysis, we propose a three-step solution pipeline including adaptive global color mapping, local enhancement and highlight generation. Moreover, the above analysis inspires us to present a lightweight network that utilizes global statistics as guidance to conduct image-adaptive color mapping. In addition, we construct a dataset using HDR videos in HDR10 standard, named HDRTV1K, and select five metrics to evaluate the results of SDRTV-to-HDRTV algorithms. Furthermore, our final results achieve state-of-the-art performance in quantitative comparisons and visual quality. The code and dataset are available at https://***/ chxy95/HDRTVNet. Copyright © 2021, The Authors. All rights reserved.

关键词： Color