计算机视觉论文-2021-07-23

本专栏是计算机视觉方向论文收集积累，时间：2021年7月23日，来源：paper digest

欢迎关注原创公众号 【计算机视觉联盟】，回复 【西瓜书手推笔记】 可获取我的机器学习纯手推笔记！

直达笔记地址：机器学习手推笔记（GitHub地址）

1, TITLE: A Public Ground-Truth Dataset for Handwritten Circuit Diagram Images
AUTHORS: Felix Thoma ; Johannes Bayer ; Yakun Li
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This paper presents such an image set along with annotations.

2, TITLE: Semantic Text-to-Face GAN -ST^2FG
AUTHORS: Manan Oza ; Sukalpa Chanda ; David Doermann
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: In this paper, we present a novel approach to generate facial images from semantic text descriptions.

3, TITLE: Triplet Is All You Need with Random Mappings for Unsupervised Visual Representation Learning
AUTHORS: WENBIN LI et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we argue that negative pairs are still necessary but one is sufficient, i.e., triplet is all you need.

4, TITLE: Correspondence-Free Point Cloud Registration with SO(3)-Equivariant Implicit Shape Representations
AUTHORS: Minghan Zhu ; Maani Ghaffari ; Huei Peng
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: This paper proposes a correspondence-free method for point cloud rotational registration.

5, TITLE: Reading Race: AI Recognises Patient's Racial Identity In Medical Images
AUTHORS: IMON BANERJEE et. al.
CATEGORY: cs.CV [cs.CV, cs.CY, eess.IV, 68-XX, I.2]
HIGHLIGHT: Interpretation: We emphasize that model ability to predict self-reported race is itself not the issue of importance.

6, TITLE: AnonySIGN: Novel Human Appearance Synthesis for Sign Language Video Anonymisation
AUTHORS: Ben Saunders ; Necati Cihan Camgoz ; Richard Bowden
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we formally introduce the task of Sign Language Video Anonymisation (SLVA) as an automatic method to anonymise the visual appearance of a sign language video whilst retaining the meaning of the original sign language sequence.

7, TITLE: 3D Shape Generation with Grid-based Implicit Functions
AUTHORS: Moritz Ibing ; Isaak Lim ; Leif Kobbelt
CATEGORY: cs.CV [cs.CV, cs.GR, cs.LG]
HIGHLIGHT: To remedy these issues, we propose to train the GAN on grids (i.e. each cell covers a part of a shape).

8, TITLE: EAN: Event Adaptive Network for Enhanced Action Recognition
AUTHORS: YUAN TIAN et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a unified action recognition framework to investigate the dynamic nature of video content by introducing the following designs.

9, TITLE: MFGNet: Dynamic Modality-Aware Filter Generation for RGB-T Tracking
AUTHORS: XIAO WANG et. al.
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: Different from these works, we propose a new dynamic modality-aware filter generation module (named MFGNet) to boost the message communication between visible and thermal data by adaptively adjusting the convolutional kernels for various input images in practical tracking.

10, TITLE: DOVE: Learning Deformable 3D Objects By Watching Videos
AUTHORS: Shangzhe Wu ; Tomas Jakab ; Christian Rupprecht ; Andrea Vedaldi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose to use monocular videos, which naturally provide correspondences across time, allowing us to learn 3D shapes of deformable object categories without explicit keypoints or template shapes.

11, TITLE: Deep 3D-CNN for Depression Diagnosis with Facial Video Recording of Self-Rating Depression Scale Questionnaire
AUTHORS: Wanqing Xie ; Lizhong Liang ; Yao Lu ; Hui Luo ; Xiaofeng Liu
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We use a new dataset of 200 participants to demonstrate the validity of self-rating questionnaires and their accompanying question-by-question video recordings in this study.

12, TITLE: PoseDet: Fast Multi-Person Pose Estimation Using Pose Embedding
AUTHORS: CHENYU TIAN et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This simple framework achieves an unprecedented speed and a competitive accuracy on the COCO benchmark compared with state-of-the-art methods.

13, TITLE: Query2Label: A Simple Transformer Way to Multi-Label Classification
AUTHORS: Shilong Liu ; Lei Zhang ; Xiao Yang ; Hang Su ; Jun Zhu
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This paper presents a simple and effective approach to solving the multi-label classification problem.

14, TITLE: External-Memory Networks for Low-Shot Learning of Targets in Forward-Looking-Sonar Imagery
AUTHORS: Isaac J. Sledge ; Christopher D. Toole ; Joseph A. Maestri ; Jose C. Principe
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: We propose a memory-based framework for real-time, data-efficient target analysis in forward-looking-sonar (FLS) imagery.

15, TITLE: Structure Destruction and Content Combination for Face Anti-Spoofing
AUTHORS: KE-YUE ZHANG et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose Structure Destruction Module and Content Combination Module to address these two imitations separately.

16, TITLE: CogSense: A Cognitively Inspired Framework for Perception Adaptation
AUTHORS: Hyukseong Kwon ; Amir Rahimi ; Kevin G. Lee ; Amit Agarwal ; Rajan Bhattacharyya
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This paper proposes the CogSense system, which is inspired by sense-making cognition and perception in the mammalian brain to perform perception error detection and perception parameter adaptation using probabilistic signal temporal logic.

17, TITLE: Geometric Data Augmentation Based on Feature Map Ensemble
AUTHORS: Takashi Shibata ; Masayuki Tanaka ; Masatoshi Okutomi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a novel CNN architecture that can improve the robustness against geometric transformations without modifying the existing backbones of their CNNs.

18, TITLE: DeepScale: An Online Frame Size Adaptation Framework to Accelerate Visual Multi-object Tracking
AUTHORS: Keivan Nalaie ; Rong Zheng
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Recognizing the effects of frame sizes on tracking performance, we propose DeepScale, a model agnostic frame size selection approach that operates on top of existing fully convolutional network-based trackers to accelerate tracking throughput.

19, TITLE: Copy and Paste Method Based on Pose for ReID
AUTHORS: Cheng Yang
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: To solve this problem, this paper proposes a simple and effective way to generate images in some new scenario, which is named Copy and Paste method based on Pose(CPP).

20, TITLE: Adaptive Dilated Convolution For Human Pose Estimation
AUTHORS: ZHENGXIONG LUO et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Towards these issues, we propose an adaptive dilated convolution (ADC).

21, TITLE: Abstract Reasoning Via Logic-guided Generation
AUTHORS: Sihyun Yu ; Sangwoo Mo ; Sungsoo Ahn ; Jinwoo Shin
CATEGORY: cs.LG [cs.LG, cs.AI, cs.CV, cs.LO]
HIGHLIGHT: To this end, we propose logic-guided generation (LoGe), a novel generative DNN framework that reduces abstract reasoning as an optimization problem in propositional logic.

22, TITLE: Improve Learning from Crowds Via Generative Augmentation
AUTHORS: Zhendong Chu ; Hongning Wang
CATEGORY: cs.LG [cs.LG, cs.CV, cs.HC]
HIGHLIGHT: In this paper, we study how to handle sparsity in crowdsourced data using data augmentation.

23, TITLE: Unsupervised Detection of Adversarial Examples with Model Explanations
AUTHORS: Gihyuk Ko ; Gyumin Lim
CATEGORY: cs.LG [cs.LG, cs.CR, cs.CV]
HIGHLIGHT: In this paper, we propose a simple yet effective method to detect adversarial examples, using methods developed to explain the model's behavior.

24, TITLE: Rethinking Trajectory Forecasting Evaluation
AUTHORS: Boris Ivanovic ; Marco Pavone
CATEGORY: cs.RO [cs.RO, cs.CV, cs.LG, cs.SY, eess.SY]
HIGHLIGHT: In this work, we take a step back and critically evaluate current trajectory forecasting metrics, proposing task-aware metrics as a better measure of performance in systems where prediction is being deployed.

25, TITLE: Self-transfer Learning Via Patches: A Prostate Cancer Triage Approach Based on Bi-parametric MRI
AUTHORS: Alvaro Fernandez-Quilez ; Trygve Eftest�l ; Morten Goodwin ; Svein Reidar Kjosavik ; Ketil Oppedal
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: In this paper, we present a patch-based pre-training strategy to distinguish between cS and ncS lesions which exploit the region of interest (ROI) of the patched source domain to efficiently train a classifier in the full-slice target domain which does not require annotations by making use of transfer learning (TL).

26, TITLE: Fristograms: Revealing and Exploiting Light Field Internals
AUTHORS: Thorsten Herfet ; Kelvin Chelli ; Tobias Lange ; Robin Kremer
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: The primary idea in this paper is to establish a relation between the capturing setup and the rays of the LF.

27, TITLE: Segmentation of Cardiac Structures Via Successive Subspace Learning with Saab Transform from Cine MRI
AUTHORS: XIAOFENG LIU et. al.
CATEGORY: eess.IV [eess.IV, cs.CV, cs.LG]
HIGHLIGHT: In this work, to address the limitations, we propose a lightweight and interpretable machine learning model, successive subspace learning with the subspace approximation with adjusted bias (Saab) transform, for accurate and efficient segmentation from cine MRI.

28, TITLE: Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data
AUTHORS: Xintao Wang ; Liangbin Xie ; Chao Dong ; Ying Shan
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: In this work, we extend the powerful ESRGAN to a practical restoration application (namely, Real-ESRGAN), which is trained with pure synthetic data.

29, TITLE: A Deep Learning-based Quality Assessment and Segmentation System with A Large-scale Benchmark Dataset for Optical Coherence Tomographic Angiography Image
AUTHORS: YUFEI WANG et. al.
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: To address these issues, we develop an automated computer-aided OCTA image processing system using deep neural networks as the classifier and segmentor to help ophthalmologists in clinical diagnosis and research.

30, TITLE: MmPose-NLP: A Natural Language Processing Approach to Precise Skeletal Pose Estimation Using MmWave Radars
AUTHORS: Arindam Sengupta ; Siyang Cao
CATEGORY: eess.SP [eess.SP, cs.CV]
HIGHLIGHT: In this paper we presented mmPose-NLP, a novel Natural Language Processing (NLP) inspired Sequence-to-Sequence (Seq2Seq) skeletal key-point estimator using millimeter-wave (mmWave) radar data.