MVA2023 Small Object Detection Challenge for Spotting Birds: Dataset, Methods, and Results (2023-07-18T00:00:00.000000Z)

TL;DR

A new SOD dataset consisting of 39,070 images including 137,121 bird instances, which is called the Small Object Detection for Spotting Birds (SOD4SB) dataset is proposed, and the award-winning methods are briefly introduced.

Abstract

Small Object Detection (SOD) is an important machine vision topic because (i) a variety of real-world applications require object detection for distant objects and (ii) SOD is a challenging task due to the noisy, blurred, and less-informative image appearances of small objects. This paper proposes a new SOD dataset consisting of 39,070 images including 137,121 bird instances, which is called the Small Object Detection for Spotting Birds (SOD4SB) dataset. The detail of the challenge with the SOD4SB dataset 1 is introduced in this paper. In total, 223 participants joined this challenge. This paper briefly introduces the award-winning methods. The dataset 2, the baseline code 3, and the website for evaluation on the public testset 4 are publicly available.

Authors

Chien-Yao Wang

4 papers

Yosuke Shinya

2 papers

Yuki Kondo

4 papers

TL;DR

Abstract

Authors

References67 items

Ensemble Fusion for Small Object Detection

BandRe: Rethinking Band-Pass Filters for Scale-Wise Object Detection Evaluation

AirBirds: A Large-scale Challenging Dataset for Bird Strike Prevention in Real-world Airports

EVA: Exploring the Limits of Masked Visual Representation Learning at Scale

InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutions

Tracking Small and Fast Moving Objects: A Benchmark

A Guide to Image- and Video-Based Small Object Detection Using Deep Learning: Case Study of Maritime Surveillance

Detecting tiny objects in aerial images: A normalized Wasserstein distance and a new benchmark

Deep Learning-based Small Object Detection: A Survey

Slicing Aided Hyper Inference and Fine-Tuning for Small Object Detection

Tiny Object Tracking: A Large-Scale Dataset and a Baseline

A Survey of the Four Pillars for Small Object Detection: Multiscale Representation, Contextual Information, Super-Resolution, and Region Proposal

Drone-vs-Bird Detection Challenge at IEEE AVSS2021

A Normalized Gaussian Wasserstein Distance for Tiny Object Detection

Distant Bird Detection for Safe Drone Flight and Its Dataset

Automated Counting Wild Birds on UAV Image Using Deep Learning

USB: Universal-Scale Object Detection Benchmark

Correlation-based ConvNet for Small Object Detection in Videos

Tiny Object Detection in Aerial Images

Simple Copy-Paste is a Strong Data Augmentation Method for Instance Segmentation

Scaled-YOLOv4: Scaling Cross Stage Partial Network

The 1st Tiny Object Detection Challenge: Methods and Results

VarifocalNet: An IoU-aware Dense Object Detector

Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection

DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution

End-to-End Object Detection with Transformers

An Evaluation of Deep Learning Methods for Small Object Detection

Scale Match for Tiny Person Detection

EfficientDet: Scalable and Efficient Object Detection

WiderPerson: A Diverse Dataset for Dense Pedestrian Detection in the Wild

iSAID: A Large-scale Dataset for Instance Segmentation in Aerial Images

CenterNet: Keypoint Triplets for Object Detection

Objects as Points

Cascade R-CNN: Delving Into High Quality Object Detection

mixup: Beyond Empirical Risk Minimization

Distress feeding of depredatory birds in Sunflower and Sorghum protected by bioacoustics

Bird detection and species classification with time‐lapse images around a wind farm: Dataset construction and evaluation

A deep learning approach to traffic lights: Detection, tracking, and classification

R-CNN for Small Object Detection

Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network

Traffic-Sign Detection and Classification in the Wild

Deep Residual Learning for Image Recognition

Construction of a bird image dataset for ecological investigations

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks

U-Net: Convolutional Networks for Biomedical Image Segmentation

Microsoft COCO: Common Objects in Context

Autonomous Unmanned Aerial Vehicle System for Controlling Pest Bird Population in Vineyards

Bird migration studies and potential collision risk with offshore wind turbines

HARKing: Hypothesizing After the Results are Known

THE USE OF DISTRESS CALLS TO REPEL NIGHT HERONS (NYCTICORAX NYCTICORAX) FROM FISH PONDS

CodaLab Competitions An Open Source Platform to Organize Scientiﬁc Challenges

CodaLab Competitions: An Open Source Platform to Organize Scientific Challenges

Swin Transformer: Hierarchical Vision Transformer using Shifted Windows

Weighted boxes fusion: Ensembling boxes from different object detection models

“Birds ﬂying dataset.”

STDnet: A ConvNet for Small Target Detection

Cascade R-CNN: Delving Authorized licensed use limited to the terms of the applicable license agreement with IEEE. Restrictions apply. into high quality object detection

experimental, theoretical and numerical investigation

International Journal of Computer Vision manuscript No. (will be inserted by the editor) Efficiently Scaling Up Crowdsourced Video Annotation A Set of Best Practices for High Quality, Economical Video Labeling

HARKing: Hypothesizing After the Results are Known

Estimating Bird Damage from Damage Incidence in Wine Grape Vineyards

“YOLO by Ultr-alytics.”

Ieee Transactions on Pattern Analysis and Machine Intelligence 1 80 Million Tiny Images: a Large Dataset for Non-parametric Object and Scene Recognition

International Journal of Computer Vision manuscript No. (will be inserted by the editor) The PASCAL Visual Object Classes (VOC) Challenge

“Base-line code for SOD4SB by IIM-TTIJ.”

Bird strike: an experimental Authorized licensed use limited to the terms of the

COCO API

Field of Study

Journal Information

Name

Page