点击我爱计算机视觉标星，更快获取CVML新技术

---

本图来自第14篇论文

本文总结了过去一周的开源计算机视觉相关代码，有好几篇来自顶会NeurIPS 2019、ICCV 2019 等。

涉及方向众多，包括模型迁移、实时目标检测、目标预测、去雨质量评价、人体运动迁移、知识蒸馏、强化学习等。

通过属性映射研究深度模型知识的可迁移性

Deep Model Transferability from Attribution Maps

Jie Song, Yixin Chen, Xinchao Wang, Chengchao Shen, Mingli Song

NeurIPS 2019

https://arxiv.org/abs/1909.11902v1

https://github.com/zju-vipa/TransferbilityFromAttributionMaps

多目标位置预测

Multiple Object Forecasting: Predicting Future Object Locations in Diverse Environments

Olly Styles, Tanaya Guha, Victor Sanchez

WACV 2020

https://arxiv.org/abs/1909.11944v1

https://github.com/olly-styles/Multiple-Object-Forecasting

对真实下雨图像的去雨质量评价，主观和客观方法

Subjective and Objective De-raining Quality Assessment Towards Authentic Rain Image

Qingbo Wu, Lei Wang, King N. Ngan, Hongliang Li, Fanman Meng

https://arxiv.org/abs/1909.11983v1

https://github.com/wqb-uestc

人体运动模仿、表观迁移和新视图合成的统一框架

Liquid Warping GAN: A Unified Framework for Human Motion Imitation, Appearance Transfer and Novel View Synthesis

Wen Liu, Zhixin Piao, Jie Min, Wenhan Luo, Lin Ma, Shenghua Gao

ICCV 2019

https://arxiv.org/abs/1909.12224v1

https://svip-lab.github.io/project/impersonator.html

基于信息多重蒸馏网络的轻量图像超分辨率

Lightweight Image Super-Resolution with Information Multi-distillation Network

Zheng Hui, Xinbo Gao, Yunchu Yang, Xiumei Wang

ACM Multimedia 2019

https://arxiv.org/abs/1909.11856v1

https://github.com/Zheng222/IMDN

隐式语义数据增强，提高了ResNets 和 DenseNets 等网络在各种数据集比如 CIFAR-10, CIFAR-100 and ImageNet上的泛化性.

Implicit Semantic Data Augmentation for Deep Networks

Yulin Wang, Xuran Pan, Shiji Song, Hong Zhang, Cheng Wu, Gao Huang

NeurIPS 2019

https://arxiv.org/abs/1909.12220v1

https://github.com/blackfeather-wang/ISDA-for-Deep-Networks

针对CNN的降低内存使用的压缩感知训练方法

CAT: Compression-Aware Training for bandwidth reduction

Chaim Baskin, Brian Chmiel, Evgenii Zheltonozhskii, Ron Banner, Alex M. Bronstein, Avi Mendelson

https://arxiv.org/abs/1909.11481v1

https://github.com/CAT-teams/CAT

知识蒸馏的“无师自通”方法

Revisit Knowledge Distillation: a Teacher-free Framework

Li Yuan, Francis E.H.Tay, Guilin Li, Tao Wang, Jiashi Feng

https://arxiv.org/abs/1909.11723v1

https://github.com/yuanli2333/Teacher-free-Knowledge-Distillation

多步视觉任务的强化学习

"Good Robot!": Efficient Reinforcement Learning for Multi-Step Visual Tasks via Reward Shaping

Andrew Hundt, Benjamin Killeen, Heeyeon Kwon, Chris Paxton, Gregory D. Hager

https://arxiv.org/abs/1909.11730v1

https://github.com/jhu-lcsr/good_robot

统一的视觉语言预训练，针对图像描述与问答

Unified Vision-Language Pre-Training for Image Captioning and VQA

Luowei Zhou, Hamid Palangi, Lei Zhang, Houdong Hu, Jason J. Corso, Jianfeng Gao

https://arxiv.org/abs/1909.11059v1

https://github.com/LuoweiZhou/VLP

PolSAR图像分类

PolSAR Image Classification Based on Dilated Convolution and Pixel-Refining Parallel Mapping network in the Complex Domain

Xiao Dongling, Liu Chang

https://arxiv.org/abs/1909.10783v1

https://github.com/PROoshio/CRPM-Net

联合人头部和身体关系学习的检测方法

Relational Learning for Joint Head and Human Detection

Cheng Chi, Shifeng Zhang, Junliang Xing, Zhen Lei, Stan Z. Li, Xudong Zou

https://arxiv.org/abs/1909.10674v1

（代码将开源，还未公布地址）

巧合的是，本周还有另一篇来自旷视的人头和身体结合的行人检测方法：

Double Anchor R-CNN for Human Detection in a Crowd

Kevin Zhang, Feng Xiong, Peize Sun, Li Hu, Boxun Li, Gang Yu

https://arxiv.org/abs/1909.09998v1

快速精确的卷积目标检测，用于实时嵌入式平台

Fast and Accurate Convolutional Object Detectors for Real-time Embedded Platforms

Min-Kook Choi, Jaehyung Park, Heechul Jung, Jinhee Lee, Soo-Heang Eo

https://arxiv.org/abs/1909.10798v1

https://github.com/mkchoi-0323/modified_refinedet

用于目标检测平衡训练的生成正样本包围框的方法

Generating Positive Bounding Boxes for Balanced Training of Object Detectors

Kemal Oksuz, Baris Can Cam, Emre Akbas, Sinan Kalkan

https://arxiv.org/abs/1909.09777v1

（代码将开源，还未公布地址）

3D点云上进行有效图卷积的球形卷积核

Spherical Kernel for Efficient Graph Convolution on 3D Point Clouds

Huan Lei, Naveed Akhtar, Ajmal Mian

https://arxiv.org/abs/1909.09287v1

https://github.com/hlei-ziyan/SPH3D-GCN

从文本图像中提取实体

EATEN: Entity-aware Attention for Single Shot Visual Text Extraction

He guo, Xiameng Qin, Jiaming Liu, Junyu Han, Jingtuo Liu, Errui Ding

https://arxiv.org/abs/1909.09380v1

https://github.com/beacandler/EATEN

---

CV细分方向交流群

52CV已经建立多个CV专业交流群，包括：目标跟踪、目标检测、语义分割、姿态估计、人脸识别检测、医学影像处理、超分辨率、神经架构搜索、GAN、强化学习等，扫码添加CV君拉你入群，如已经为CV君其他账号好友请直接私信，

（请务必注明相关方向，比如：目标检测）

喜欢在QQ交流的童鞋，可以加52CV官方QQ群：805388940。

（不会时时在线，如果没能及时通过验证还请见谅）

---

长按关注我爱计算机视觉