Reduced Focal Loss: 1st Place Solution to xView object detection in Satellite Imagery

This paper describes our approach to the DIUx xView 2018 Detection Challenge [1]. This challenge focuses on a new satellite imagery dataset. The dataset contains 60 object classes that are highly imbalanced. Due to the imbalanced nature of the dataset, the training process becomes significantly more...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2019-04
Main Authors: Sergievskiy, Nikolay, Ponamarev, Alexander
Format: Article
Language:English
Subjects:
Datasets
Image detection
Object recognition
Satellite imagery
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper describes our approach to the DIUx xView 2018 Detection Challenge [1]. This challenge focuses on a new satellite imagery dataset. The dataset contains 60 object classes that are highly imbalanced. Due to the imbalanced nature of the dataset, the training process becomes significantly more challenging. To address this problem, we introduce a novel Reduced Focal Loss function, which brought us 1st place in the DIUx xView 2018 Detection Challenge.
ISSN:2331-8422