Skip to content

Latest commit

 

History

History
83 lines (64 loc) · 4.62 KB

COCO and Pascal VOC.md

File metadata and controls

83 lines (64 loc) · 4.62 KB

COCO and Pascal VOC

Browse > Computer Vision > Object Detection - [Link]

COCO

Introduction

COCO is an image dataset designed to spur object detection research with a focus on detecting objects in context. The annotations include instance segmentations for object belonging to 80 categories, stuff segmentations for 91 categories, keypoint annotations for person instances, and five image captions per image.

COCO is a large-scale object detection, segmentation, and captioning dataset. COCO has several features:

Object segmentation
Recognition in context
Superpixel stuff segmentation
330K images (>200K labeled)
1.5 million object instances
80 object categories
91 stuff categories
5 captions per image
250,000 people with keypoints

COCO2017 Data


VOC2007

Introduction

The goal of this challenge is to recognize objects from a number of visual object classes in realistic scenes (i.e. not pre-segmented objects). It is fundamentally a supervised learning learning problem in that a training set of labelled images is provided. The twenty object classes that have been selected are:

Statistics
20 classes:
Person: person
Animal: bird, cat, cow, dog, horse, sheep
Vehicle: aeroplane, bicycle, boat, bus, car, motorbike, train
Indoor: bottle, chair, dining table, potted plant, sofa, tv/monitor
Train/validation/test: 9,963 images containing 24,640 annotated objects.
New developments 
Number of classes increased from 10 to 20
Segmentation taster introduced
Person layout taster introduced
Truncation flag added to annotations
Evaluation measure for the classification challenge changed to Average Precision. Previously it had been ROC-AUC.

VOC2007 Data

The annotated test data for the VOC challenge 2007 is now available:

VOC2012 Data

The development kit is now available:

References

ECCV 2018 Joint COCO and Mapillary Recognition - [Link]

The PASCAL Visual Object Classes Challenge 2007 - [Link]

Object detection: speed and accuracy comparison - [Link]