AdaptIS: Adaptive Instance Selection Network

2019-09-19 12:58:07

Paperhttps://arxiv.org/pdf/1909.07829.pdf

Code (MXNet)https://github.com/saic-vul/adaptis

Pretrained model for ToyV1: https://drive.google.com/open?id=1IuJUh0JvbKYILBxCeO2h6U4LG-9DoTHi
Pretrained model for ToyV2: https://drive.google.com/open?id=1RxepfpJF5gRpRNYu1urdV748suF3TL5k

Related Paper:

Panoptic Segmentation, Alexander Kirillov, Kaiming He, Ross Girshick, Carsten Rother, Piotr Dollar

Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization [Paper] [Code] [Blog]

An intriguing failing of convolutional neural networks and the CoordConv solution [Paper] [Code] [Blog]

1. Background and Motivation:

本文提出一种新的分割方式,即:给出一个 BBox,该方法可以将该位置的物体分割出来,而不是全部分割出来。示意图如下所示:

本文所提出方法的名称为:AdaptIS,不依赖于 bounding box proposal。而是直接优化目标分割精度。给定一张图像 I 和 一个固定的 point proposal (x, y),作者直接优化目标损失函数。我们利用一个 pixel-wise loss 来计算 AdaptIS 预测 和  target object 的 mask。

AdaptIS: Adaptive Instance Selection Network的更多相关文章

  1. openstack中虚拟机和其网络的联系方法 instance and network

    instance和network的连接关系: 在neutron数据库中,ports存储了instance和端口的关系,其中device_id表示了instance的id,在nova数据库中的insta ...

  2. 论文笔记——An online EEG-based brain-computer interface for controlling hand grasp using an adaptive probabilistic neural network(10年被引用66次)

    题目:利用自适应概率网络设计一种在线脑机接口楼方法控制手部抓握 概要:这篇文章提出了一种新的脑机接口方法,控制手部,系列手部抓握动作和张开在虚拟现实环境中.这篇文章希望在现实生活中利用脑机接口技术控制 ...

  3. Two-Stream Adaptive Graph Convolutional Network for Skeleton-Based Action Recognition

    Two-Stream Adaptive Graph Convolutional Network for Skeleton-Based Action Recognition 摘要 基于骨架的动作识别因为 ...

  4. Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization

    Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization 2019-10-10 10:50:19 Paper ...

  5. 论文《Entity Linking with Effective Acronym Expansion, Instance Selection and Topic Modeling》

    Entity Linking with Effective Acronym Expansion, Instance Selection and Topic Modeling 一.主要贡献 1. pro ...

  6. 【论文速读】Fangfang Wang_CVPR2018_Geometry-Aware Scene Text Detection With Instance Transformation Network

    Han Hu--[ICCV2017]WordSup_Exploiting Word Annotations for Character based Text Detection 作者和代码 caffe ...

  7. An intriguing failing of convolutional neural networks and the CoordConv solution

    An intriguing failing of convolutional neural networks and the CoordConv solution NeurIPS 2018 2019- ...

  8. [paper reading] C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection CVPR2019

    MIL陷入局部最优,检测到局部,无法完整的检测到物体.将instance划分为空间相关和类别相关的子集.在这些子集中定义一系列平滑的损失近似代替原损失函数,优化这些平滑损失. C-MIL learns ...

  9. {ICIP2014}{收录论文列表}

    This article come from HEREARS-L1: Learning Tuesday 10:30–12:30; Oral Session; Room: Leonard de Vinc ...

随机推荐

  1. Implement CGLIB in ABAP

    What is cglib?A Byte Code Generation Library which is high level API to generate and transform Java ...

  2. 小程序中的数据请求sessionid,保持登陆状态。

    版权声明:本文为CSDN博主「weixin_43964779」的原创文章,遵循 CC 4.0 BY-SA 版权协议,转载请附上原文出处链接及本声明.原文链接:https://blog.csdn.net ...

  3. Linux一些服务的启动命令

    http:服务service httpd start 启动 service httpd restart 重新启动 service httpd stop 停止服务 启动ssh服务 # /etc/init ...

  4. 有意义的单词分割——经典dfs题目

    680. 分割字符串 中文 English 给一个字符串,你可以选择在一个字符或两个相邻字符之后拆分字符串,使字符串由仅一个字符或两个字符组成,输出所有可能的结果 样例 样例1 输入: "1 ...

  5. Json在序列化注意问题

    Java中的Json序列化,不容忽视的getter 问题重现 public class AjaxJson { private boolean success; private String msg; ...

  6. Spark数据倾斜解决方案(转)

    本文转发自技术世界,原文链接 http://www.jasongj.com/spark/skew/ Spark性能优化之道——解决Spark数据倾斜(Data Skew)的N种姿势  发表于 2017 ...

  7. 字符串翻转(C++)

    1.字符串原地翻转,"abc"->"cba": int str_reverse(string &str,int first,int last) { ...

  8. LeetCode 855. Exam Room

    原题链接在这里:https://leetcode.com/problems/exam-room/ 题目: In an exam room, there are N seats in a single ...

  9. Tensorflow细节-P212-循环神经网络

    本节的循环神经网络一图足以说明 import numpy as np X = [1, 2] state = [0.0, 0.0] # 定义RNN的参数 # 以下两个本来是像这样分开的,但是在运算时合并 ...

  10. chef test-kitchen Could not load the 'vagrant' driver from the load path 问题解决

    今天使用chef 的kitchen,运行kitchen list 发现了如下错误: >>>>>> ------Exception------- >>&g ...