计算机视觉顶级赛事ICCV 2017在威尼斯悄然落幕，期间中国团队在物体检测、人体关键点检测等竞争激烈的比赛中击败了谷歌、微软、Facebook等国际巨头AI实验室。

COCO+Places 2017

MS COCO是一个已经举办了三年，在业内颇有名气的比赛。今年的MS COCO共有四个项目，包括物体检测、物体分割、人体关键点检测和场景分割。

和MS COCO联合公布结果的Places今年还是第一届，由MIT和CMU牵头，Places 2017的挑战主要有三个任务：场景分割（scene parsing）、物体分割（instance segmentation）、边缘检测（semantic boundary detection）。

Places官网地址：http://placeschallenge.csail.mit.edu/

结果公布网址：http://placeschallenge.csail.mit.edu/results_challenge.html

数据集地址：http://groups.csail.mit.edu/vision/datasets/ADE20K/

原文：

Places Challenge 2017: Deep Scene Understanding is held jointly with COCO Challenge at ICCV'17. Scene understanding is one of the hallmark tasks of computer vision, allowing the definition of a context for object recognition. The goal of the Places Challenge is to stimulate the computer vision community to develop new algorithms and models that improve the state of the arts in visual scene understanding. Winners will be invited to present at Joint COCO and Places Challenge Workshop at ICCV 2017.

There are three tasks in Places Challenge 2017: Scene Parsing, Scene Instance Segmentation, and Semantic Boundary Detection. The data for all the three tasks are from the fully annotated image dataset ADE20K, there are 20K images for training, 2K images for validation, and 3K images for testing. Teams could particpate in one or two or three of the tasks. The details for each task are listed below:

COCO挑战赛

COCO是一个图像数据集，被设计用来推动物体检测研究，特别是检测上下文中的物体。其中提供的注释包括80个分类的物体像素级分割，人体实例的关键点注释，91个类别的背景语义分割。

COCO官网地址：

http://cocodataset.org/

https://places-coco2017.github.io/

大赛具体包括：COCO检测挑战

COCO 2017检测挑战赛已在推动物体检测领域的进步。参赛队伍要在两类物体检测挑战中竞争：使用包围盒（bounding box）输出或者物体分割输出。

COCO关键点挑战

这项挑战需要在复杂环境下对人体关键点进行定位。这项挑战需要在检测出人体的同时，对关键点进行定位标注。

COCO背景语义分割挑战

今年的挑战中，已经给出人、汽车、大象等物体的分类，所以重点主要在背景分类的部分，例如草坪、墙壁、天空等。

1. Overview

The goal of the joint COCO and Places Challenge is to study object recognition in the context of scene understanding.

2. COCO Challenges

COCO is an image dataset designed to spur object detection research with a focus on detecting objects in context. The annotations include pixel-level segmentation of object belonging to 80 categories, keypoint annotations for person instances, stuff segmentations for 91 categories, and five image captions per image. The specific tracks in the COCO 2017 Challenges are (1) object detection with bounding boxes and segmentation masks, (2) joint detection and person keypoint estimation, and (3) stuff segmentation. We describe each next.

2.1. COCO Detection Challenge

The COCO 2017 Detection Challenge is designed to push the state of the art in object detection forward. Teams are encouraged to compete in either (or both) of two object detection challenges: using bounding box output or object segmentation output. For full details of this task please see the COCO Detection Challenge page.

2.2. COCO Keypoint Challenge

The COCO 2017 Keypoint Challenge requires localization of person keypoints in challenging, uncontrolled conditions. The keypoint challenge involves simultaneously detecting people and localizing their keypoints (person locations are not given at test time). For full details of this task please see the COCO Keypoints Challenge page.

2.3. COCO Stuff Challenge

The COCO 2017 Stuff Segmentation Challenge is designed to push the state of the art in semantic segmentation of stuff classes. Whereas the COCO 2017 Detection Challenge addresses thing classes (person, car, elephant), this challenge focuses on stuff classes (grass, wall, sky). For full details of this task please see the COCO Stuff Challenge page.

3. Places Challenges

The Places Challenge will host three tracks meant to complement the COCO Challenges. The data for the 2017 Places Challenge is from the pixel-wise annotated image dataset ADE20K, in which there are 20K images for training, 2K validation images, and 3K testing images. The three specific tracks in the Places Challenge 2017 are: (1) scene parsing, (2) instance segmentation, and (3) semantic boundary detection. See the Places Challenge Page for detailed information.

本次大赛结果

中国AI创业公司旷视科技（Face++）在MS COCO物体检测、人体关键点检测，以及Places物体分割三项比赛中击败微软、谷歌、Facebook等对手，夺得了第一名。

旷视科技获COCO物体检测、人体关键点检测冠军；UCenter获COCO物体分割冠军

而在MS COCO物体分割检测中，由北京大学和香港中文大学联合组成的UCenter队（也可以理解为商汤科技队）夺得冠军，旷视科技（Face++）团队获得了第二名。

Places场景分割挑战赛的冠军由中科院自动化所和京东联合建立的CASIA_IVA_JD队拿下，第二名是今日头条的WinterIsComing队。