Endoscopy artifact detection (EAD 2019) challenge dataset

@article{Ali2019EndoscopyAD,
  title={Endoscopy artifact detection (EAD 2019) challenge dataset},
  author={Sharib Ali and Felix Y Zhou and Christian Daul and Barbara Braden and Adam Bailey and Stefano Realdon and James Edward East and Georges Wagni{\`e}res and Victor Loschenov and Enrico Grisan and Walter Blondel and J. Rittscher},
  journal={ArXiv},
  year={2019},
  volume={abs/1905.03209}
}
Endoscopic artifacts are a core challenge in facilitating the diagnosis and treatment of diseases in hollow organs. Precise detection of specific artifacts like pixel saturations, motion blur, specular reflections, bubbles and debris is essential for high-quality frame restoration and is crucial for realizing reliable computer-assisted tools for improved patient care. At present most videos in endoscopy are currently not analyzed due to the abundant presence of multi-class artifacts in video… 

Figures from this paper

Artefact detection in video endoscopy using retinanet and focal loss function
TLDR
This work describes its submission to the EAD 2019 challenge to detect bounding boxes for seven classes of artefacts in endoscopy videos by generating a large derivative dataset by augmenting the original images with free-form deformations to prevent over-fitting.
An objective comparison of detection and segmentation algorithms for artefacts in clinical endoscopy
TLDR
A comprehensive analysis of the submissions to the first edition of the Endoscopy Artefact Detection challenge is presented, revealing the shortcomings of current training strategies and highlighting the need for developing new optimal metrics to accurately quantify the clinical applicability of methods.
FOCAL LOSS FOR ARTEFACT DETECTION IN MEDICAL ENDOSCOPY
TLDR
This work proposes an ensemble of deep learning object detectors to automate multi-class artefact detection in video endoscopy and achieved a mean average precision (mAP) and an average intersection-over-union (IoU) of 0.3997 on the EAD2019 test set.
Endoscopic computer vision challenges 2.0
TLDR
This year's challenge aimed to benchmark methods on larger test-set comprising of mostly video sequences as in the real-world clinical scenario, unlike previous challenges, in 2.0 focus on assessing generalizability of algorithms.
Endoscopy Artefact Detection by Deep Transfer Learning of Baseline Models
TLDR
A simple baseline could achieve state-of-the-art performance in detecting artefacts in endoscopy using faster region-based convolutional neural networks and the combination of EfficientDet-D2 with suitable data augmentation and pre-trained parameters during fine-tuning training to detect the artefacts.
Deep Encoder-decoder Networks for Artefacts Segmentation in Endoscopy Images
TLDR
The paper briefly describes the methodology adopted for the 2020 Endoscopy Artefact Detection and Segmentation (EAD2020) challenge, and how three off-theshelf deep detection networks have been optimised and evaluated on the EAD data.
Artifact Detection in Endoscopic Video with Deep Convolutional Neural Networks
TLDR
Experimental results show that the proposed objection detectors based on SSD and Faster-RCNN significantly outperform the baseline and improve the performance of two improved CNN-based methods for seven-class endoscopic artifact detection (EAD).
Endoscopic Artefact Detection with Ensemble of Deep Neural Networks and False Positive Elimination
TLDR
An endoscopic artefact detection framework based on an ensemble of deep neural networks, classagnostic non-maximum suppression, and false-positive elimination is proposed, exploiting the distinctive properties of different approaches.
DNN MODELS AND POSTPROCESSING THRESHOLDS FOR ENDOSCOPY ARTIFACT DETECTION IN PRACTICE
TLDR
This work proposed to combine the advantages of two state-of-the-art deep object detection algorithms, YOLOv3 and Mask R-CNN, and achieved improved performance by leveraging the available image segmentation annotations to aid bounding box detection.
...
...

References

SHOWING 1-5 OF 5 REFERENCES
A deep learning framework for quality assessment and restoration in video endoscopy
The VGG Image Annotator (VIA)
TLDR
This paper introduces a simple and standalone manual image annotation tool: the VGG Image Annotator, a light weight, standalone and offline software package that does not require any installation or setup and runs solely in a web browser.
The VIA Annotation Software for Images, Audio and Video
TLDR
A light weight, standalone and offline software package that does not require any installation or setup and runs solely in a web browser, the VIA software allows human annotators to define and describe spatial regions in images or video frames, and temporal segments in audio or video.
"J."
however (for it was the literal soul of the life of the Redeemer, John xv. io), is the peculiar token of fellowship with the Redeemer. That love to God (what is meant here is not God’s love to men)
A
  • Zisserman, The PASCAL Visual Object Classes Challenge 2012
  • 2012