本专栏是计算机视觉方向论文收集积累，时间：2021年3月18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: On The Whitney Extension Problem for Near Isometries and Beyond
AUTHORS: Steven B. Damelin
CATEGORY: math.CA [math.CA, cs.CV, cs.LG, math.OC, 42B37, 42B35, 30H35, 30E10, 14Q15, 53A45, 58Z05, 68P01, 49J35, 49J30, 49J10, 49J21]
HIGHLIGHT: On The Whitney Extension Problem for Near Isometries and Beyond
2, TITLE: Automatic Generation of Contrast Sets from Scene Graphs: Probing The Compositional Consistency of GQA
AUTHORS: Yonatan Bitton ; Gabriel Stanovsky ; Roy Schwartz ; Michael Elhadad
CATEGORY: cs.CL [cs.CL, cs.CV]
HIGHLIGHT: Recent works have shown that supervised models often exploit data artifacts to achieve good test scores while their performance severely degrades on samples outside their training distribution.
3, TITLE: Disentangled Cycle Consistency for Highly-realistic Virtual Try-On
AUTHORS: CHONGJIAN GE et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a Disentangled Cycle-consistency Try-On Network (DCTON).
4, TITLE: YOLOStereo3D: A Step Back to 2D for Efficient Stereo 3D Detection
AUTHORS: Yuxuan Liu ; Lujia Wang ; Ming Liu
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We incorporate knowledge and the inference structure from real-time one-stage 2D/3D object detector and introduce a light-weight stereo matching module.
5, TITLE: Large-Scale Zero-Shot Image Classification from Rich and Diverse Textual Descriptions
AUTHORS: Sebastian Bujwid ; Josephine Sullivan
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We study the impact of using rich and diverse textual descriptions of classes for zero-shot learning (ZSL) on ImageNet.
6, TITLE: An Efficient Method for The Classification of Croplands in Scarce-Label Regions
AUTHORS: Houtan Ghaffari
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: We introduce three self-supervised tasks for cropland classification.
7, TITLE: Digital Peter: Dataset, Competition and Handwriting Recognition Methods
AUTHORS: MARK POTANIN et. al.
CATEGORY: cs.CV [cs.CV, cs.AI, cs.LG, I.7.5; I.4.6]
HIGHLIGHT: This paper presents a new dataset of Peter the Great's manuscripts and describes a segmentation procedure that converts initial images of documents into the lines.
8, TITLE: Semi-Supervised Learning for Eye Image Segmentation
AUTHORS: Aayush K. Chaudhary ; Prashnna K. Gyawali ; Linwei Wang ; Jeff B. Pelz
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This work presents two semi-supervised learning frameworks to identify eye-parts by taking advantage of unlabeled images where labeled datasets are scarce.
9, TITLE: ALADIN: All Layer Adaptive Instance Normalization for Fine-grained Style Similarity
AUTHORS: DAN RUTA et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We present ALADIN (All Layer AdaIN); a novel architecture for searching images based on the similarity of their artistic style.
10, TITLE: Single Underwater Image Restoration By Contrastive Learning
AUTHORS: JUNLIN HAN et. al.
CATEGORY: cs.CV [cs.CV, eess.IV]
HIGHLIGHT: We design our method by leveraging from contrastive learning and generative adversarial networks to maximize mutual information between raw and restored images. Additionally, we release a large-scale real underwater image dataset to support both paired and unpaired training modules.
11, TITLE: Prediction-assistant Frame Super-Resolution for Video Streaming
AUTHORS: WANG SHEN et. al.
CATEGORY: cs.CV [cs.CV, eess.IV]
HIGHLIGHT: In this paper, we propose to enhance video quality using lossy frames in two situations.
12, TITLE: SPICE: Semantic Pseudo-labeling for Image Clustering
AUTHORS: Chuang Niu ; Ge Wang
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: This paper presents SPICE, a Semantic Pseudo-labeling framework for Image ClustEring.
13, TITLE: Learning Discriminative Prototypes with Dynamic Time Warping
AUTHORS: Xiaobin Chang ; Frederick Tung ; Greg Mori
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We propose Discriminative Prototype DTW (DP-DTW), a novel method to learn class-specific discriminative prototypes for temporal recognition tasks.
14, TITLE: Triplet-Watershed for Hyperspectral Image Classification
AUTHORS: Aditya Challa ; Sravan Danda ; B. S. Daya Sagar ; Laurent Najman
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this article, we propose to use a watershed classifier.
15, TITLE: ShipSRDet: An End-to-End Remote Sensing Ship Detector Using Super-Resolved Feature Representation
AUTHORS: Shitian He ; Huanxin Zou ; Yingqian Wang ; Runlin Li ; Fei Cheng
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we explore the potential benefits introduced by image SR to ship detection, and propose an end-to-end network named ShipSRDet.
16, TITLE: Temporal Cluster Matching for Change Detection of Structures from Satellite Imagery
AUTHORS: Caleb Robinson ; Anthony Ortiz ; Juan M. Lavista Ferres ; Brandon Anderson ; Daniel E. Ho
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We propose a general model, Temporal Cluster Matching (TCM), for detecting building changes in time series of remotely sensed imagery when footprint labels are only available for a single point in time.
17, TITLE: Virtual Dress Swap Using Landmark Detection
AUTHORS: Odar Zeynal ; Saber Malekzadeh
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: This paper addresses one crucial problem of buying dress online, which has not been solved yet.
18, TITLE: Revisiting The Loss Weight Adjustment in Object Detection
AUTHORS: WENXIN YU et. al.
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: By incorporating ALWA into both one-stage and two-stage object detectors, we show a consistent improvement on their performance using L1, SmoothL1 and CIoU loss, performance measures on popular object detection benchmarks including PASCAL VOC and MS COCO.
19, TITLE: You Only Look One-level Feature
AUTHORS: QIANG CHEN et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Based on the simple and efficient solution, we present You Only Look One-level Feature (YOLOF).
20, TITLE: WheatNet: A Lightweight Convolutional Neural Network for High-throughput Image-based Wheat Head Detection and Counting
AUTHORS: Saeed Khaki ; Nima Safaei ; Hieu Pham ; Lizhi Wang
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: To help mitigate this data collection bottleneck in wheat breeding, we propose a novel deep learning framework to accurately and efficiently count wheat heads to aid in the gathering of real-time data for decision making.
21, TITLE: A Comparative Study of Deep Learning Methods for Building Footprints Detection Using High Spatial Resolution Aerial Images
AUTHORS: HONGJIE HE et. al.
CATEGORY: cs.CV [cs.CV, eess.IV]
HIGHLIGHT: The paper aims at providing the whole process of building footprints extraction from high spatial resolution images using deep learning-based methods.
22, TITLE: Multi-channel Deep Supervision for Crowd Counting
AUTHORS: Bo Wei ; Mulin Chen ; Qi Wang ; Xuelong Li
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To tackle these problems, in this paper, we propose an effective network called MDSNet, which introduces a novel supervision framework called Multi-channel Deep Supervision (MDS).
23, TITLE: Co-Generation and Segmentation for Generalized Surgical Instrument Segmentation on Unlabelled Data
AUTHORS: Megha Kalia ; Tajwar Abrar Aleef ; Nassir Navab ; Septimiu E. Salcudean
CATEGORY: cs.CV [cs.CV, cs.LG, cs.RO, eess.IV]
HIGHLIGHT: In this paper, we demonstrate the limited generalizability of these methods on different datasets, including human robot-assisted surgeries.
24, TITLE: Trans-SVNet: Accurate Phase Recognition from Surgical Videos Via Hybrid Embedding Aggregation Transformer
AUTHORS: Xiaojie Gao ; Yueming Jin ; Yonghao Long ; Qi Dou ; Pheng-Ann Heng
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: In this paper, we introduce, for the first time in surgical workflow analysis, Transformer to reconsider the ignored complementary effects of spatial and temporal features for accurate surgical phase recognition.
25, TITLE: Interpretable Distance Metric Learning for Handwritten Chinese Character Recognition
AUTHORS: Boxiang Dong ; Aparna S. Varde ; Danilo Stevanovic ; Jiayin Wang ; Liang Zhao
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: In this paper, we propose an interpretable distance metric learning approach for handwritten Chinese character recognition.
26, TITLE: Quantitative Effectiveness Assessment and Role Categorization of Individual Units in Convolutional Neural Networks
AUTHORS: Yang Zhao ; Hao Zhang
CATEGORY: cs.CV [cs.CV, cs.LG, I.2.10]
HIGHLIGHT: To this end, we propose a novel method for quantitatively clarifying the status and usefulness of single unit of CNN in image classification tasks.
27, TITLE: Few-Shot Visual Grounding for Natural Human-Robot Interaction
AUTHORS: Giorgos Tziafas ; Hamidreza Kasaei
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: Towards addressing this point, we propose a software architecture that segments a target object from a crowded scene, indicated verbally by a human user.
28, TITLE: Hierarchical Random Walker Segmentation for Large Volumetric Biomedical Data
AUTHORS: Dominik Drees ; Xiaoyi Jiang
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We propose a hierarchical framework that, to the best of our knowledge, is the first attempt to overcome these restrictions for the random walker algorithm and achieves sublinear run time and constant memory complexity.
29, TITLE: The Invertible U-Net for Optical-Flow-free Video Interframe Generation
AUTHORS: Saem Park ; Donghun Han ; Nojun Kwak
CATEGORY: cs.CV [cs.CV, cs.LG, eess.IV]
HIGHLIGHT: Most conventional methods use optical flow, and various tools such as occlusion handling and object smoothing are indispensable.
30, TITLE: Bio-inspired Robustness: A Review
AUTHORS: Harshitha Machiraju ; Oh-Hyeon Choung ; Pascal Frossard ; Michael. H Herzog
CATEGORY: cs.CV [cs.CV, cs.AI, cs.LG]
HIGHLIGHT: We propose a set of criteria for proper evaluation and analyze different models according to these criteria.
31, TITLE: Pose-GNN : Camera Pose Estimation System Using Graph Neural Networks
AUTHORS: Ahmed Elmoogy ; Xiaodai Dong ; Tao Lu ; Robert Westendorp ; Kishore Reddy
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We propose a novel image based localization system using graph neural networks (GNN).
32, TITLE: Aggregated Multi-GANs for Controlled 3D Human Motion Prediction
AUTHORS: ZHENGUANG LIU et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a generalization of the human motion prediction task in which control parameters can be readily incorporated to adjust the forecasted motion.
33, TITLE: CNN Model & Tuning for Global Road Damage Detection
AUTHORS: Rahul Vishwakarma ; Ravigopal Vennelakanti
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We briefly describe the tuning strategy for the experiments conducted on two-stage Faster R-CNN with Deep Residual Network (Resnet) and Feature Pyramid Network (FPN) backbone. We assess single and multi-stage network architectures for object detection and provide a benchmark using popular state-of-the-art open-source PyTorch frameworks like Detectron2 and Yolov5.
34, TITLE: Fourier Transform of Percoll Gradients Boosts CNN Classification of Hereditary Hemolytic Anemias
AUTHORS: ARIO SADAFI et. al.
CATEGORY: cs.CV [cs.CV, eess.IV]
HIGHLIGHT: Here, we propose a method for identifying hereditary hemolytic anemias based on a standard biochemistry method, called Percoll gradient, obtained by centrifuging a patient's blood.
35, TITLE: HAMIL: Hierarchical Aggregation-Based Multi-Instance Learning for Microscopy Image Classification
AUTHORS: Yanlun Tu ; Houchao Lei ; Wei Long ; Yang Yang
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this study, we propose a hierarchical aggregation network for multi-instance learning, called HAMIL.
36, TITLE: Adversarial Attacks on Camera-LiDAR Models for 3D Car Detection
AUTHORS: Mazen Abdelfattah ; Kaiwen Yuan ; Z. Jane Wang ; Rabab Ward
CATEGORY: cs.CV [cs.CV, cs.CR, cs.GR, cs.LG]
HIGHLIGHT: We propose a universal and physically realizable adversarial attack for each type, and study and contrast their respective vulnerabilities to attacks.
37, TITLE: Meta-learning of Pooling Layers for Character Recognition
AUTHORS: Takato Otsuzuki ; Heon Song ; Seiichi Uchida ; Hideaki Hayashi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a meta-learning framework for pooling layers.
38, TITLE: Improved Deep Classwise Hashing With Centers Similarity Learning for Image Retrieval
AUTHORS: Ming Zhang ; Hong Yan
CATEGORY: cs.CV [cs.CV, cs.IR]
HIGHLIGHT: In this paper, we propose an improved deep classwise hashing, which enables hashing learning and class centers learning simultaneously.
39, TITLE: Pros and Cons of GAN Evaluation Measures: New Developments
AUTHORS: Ali Borji
CATEGORY: cs.LG [cs.LG, cs.AI, cs.CV]
HIGHLIGHT: I describe new dimensions that are becoming important in assessing models, and discuss the connection between GAN evaluation and deepfakes.
40, TITLE: Multi-Prize Lottery Ticket Hypothesis: Finding Accurate Binary Neural Networks By Pruning A Randomly Weighted Network
AUTHORS: James Diffenderfer ; Bhavya Kailkhura
CATEGORY: cs.LG [cs.LG, cs.CV]
HIGHLIGHT: In this paper, we propose (and prove) a stronger Multi-Prize Lottery Ticket Hypothesis: A sufficiently over-parameterized neural network with random weights contains several subnetworks (winning tickets) that (a) have comparable accuracy to a dense target network with learned weights (prize 1), (b) do not require any further training to achieve prize 1 (prize 2), and (c) is robust to extreme forms of quantization (i.e., binary weights and/or activation) (prize 3).
41, TITLE: Learning with Group Noise
AUTHORS: QIZHOU WANG et. al.
CATEGORY: cs.LG [cs.LG, cs.CV]
HIGHLIGHT: To overcome this issue, we propose a novel Max-Matching method for learning with group noise.
42, TITLE: Theoretical Bounds on Data Requirements for The Ray-based Classification
AUTHORS: Brian J. Weber ; Sandesh S. Kalantre ; Thomas McJunkin ; Jacob M. Taylor ; Justyna P. Zwolak
CATEGORY: cs.LG [cs.LG, cs.CV, stat.ML, 68T20, 68Q32, 68U10]
HIGHLIGHT: For the case of identifying convex shapes of different geometries, a new classification framework has recently been proposed in which the intersections of a set of one-dimensional representations, called rays, with the boundaries of the shape are used to identify the specific geometry.
43, TITLE: Training GANs with Stronger Augmentations Via Contrastive Discriminator
AUTHORS: Jongheon Jeong ; Jinwoo Shin
CATEGORY: cs.LG [cs.LG, cs.AI, cs.CV]
HIGHLIGHT: In this paper, we propose a novel way to address these questions by incorporating a recent contrastive representation learning scheme into the GAN discriminator, coined ContraD.
44, TITLE: PredRNN: A Recurrent Neural Network for Spatiotemporal Predictive Learning
AUTHORS: UNBO WANG et. al.
CATEGORY: cs.LG [cs.LG, cs.CV]
HIGHLIGHT: We provide detailed ablation studies, gradient analyses, and visualizations to verify the effectiveness of each component.
45, TITLE: Gradient Projection Memory for Continual Learning
AUTHORS: Gobinda Saha ; Isha Garg ; Kaushik Roy
CATEGORY: cs.LG [cs.LG, cs.CV]
HIGHLIGHT: In contrast, we propose a novel approach where a neural network learns new tasks by taking gradient steps in the orthogonal direction to the gradient subspaces deemed important for the past tasks.
46, TITLE: Generating Annotated Training Data for 6D Object Pose Estimation in Operational Environments with Minimal User Interaction
AUTHORS: Paul Koch ; Marian Schl�ter ; Serge Thill
CATEGORY: cs.RO [cs.RO, cs.CV, cs.LG]
HIGHLIGHT: Here, we present a proof of concept for a novel approach of autonomously generating annotated training data for 6D object pose estimation.
47, TITLE: What S in My LiDAR Odometry Toolbox?
AUTHORS: Pierre Dellenbach ; Jean-Emmanuel Deschaud ; Bastien Jacquet ; Fran�ois Goulette
CATEGORY: cs.RO [cs.RO, cs.CV]
HIGHLIGHT: In this paper, we review and organize the main 3D LiDAR odometries into distinct categories.
48, TITLE: On The Role of Images for Analyzing Claims in Social Media
AUTHORS: Gullal S. Cheema ; Sherzod Hakimov ; Eric M�ller-Budack ; Ralph Ewerth
CATEGORY: cs.SI [cs.SI, cs.CL, cs.CV]
HIGHLIGHT: In this paper, we present an empirical study on visual, textual, and multimodal models for the tasks of claim, claim check-worthiness, and conspiracy detection, all of which are related to fake news detection.
49, TITLE: Collapsible Linear Blocks for Super-Efficient Super Resolution
AUTHORS: KARTIKEYA BHARDWAJ et. al.
CATEGORY: eess.IV [eess.IV, cs.CV, cs.LG]
HIGHLIGHT: In this paper, we propose SESR, a new class of Super-Efficient Super Resolution networks that significantly improve image quality and reduce computational complexity.
50, TITLE: Colorectal Cancer Segmentation Using Atrous Convolution and Residual Enhanced UNet
AUTHORS: Nisarg A. Shah ; Divij Gupta ; Romil Lodaya ; Ujjwal Baid ; Sanjay Talbar
CATEGORY: eess.IV [eess.IV, cs.CV, cs.LG]
HIGHLIGHT: For the task at hand, we propose another CNN-based approach, which uses atrous convolutions and residual connections besides the conventional filters.
