Kvasir-Capsule, a video capsule endoscopy dataset

Smedsrud, Pia H.; Gjestang, Henrik L.; Nedrejord, Oda Olsen; Næss, Espen; Thambawita, Vajira; Hicks, Steven A.; Borgli, Hanna; Jha, Debesh; Berstad, Tor Jan Derek; Eskeland, Sigrun Losada; Lux, Mathias; Espeland, Håvard; Petlund, Andreas; Dang-Nguyen, Duc-Tien; Garcia-Ceja, Enrique; Johansen, Dag; Schmidt, Peter T.; Hammer, Hugo Lewi; Lange, Thomas de; Riegler, Michael; Halvorsen, Pål

doi:10.31219/osf.io/gr7bn

Cited by 10 publications

(11 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This requirement is supported by our dataset publication "Kvasir-Capsule, a Video Capsule Endoscopy Dataset" [5], and our provided GitHub repository 2 . Our initial experiments and research are done on another open source dataset, Hyper-Kvasir 3 [6].…”

Section: Main Contributionsmentioning

confidence: 99%

“…This objective is supported by our creation of a gastrointestinal dataset, Kvasir-Capsule [5], containing a total of 44,260 manually labeled images with bounding boxes around the respective finding, split into 13 classes for pathological findings, anatomical landmarks and quality of mucosal view. The annotation was performed by three MSc student, supervised by an expert endoscopist with many years in the field.…”

Section: Requirementmentioning

confidence: 99%

“…In Section 3.1.2 in the next chapter, we will introduce how we gathered data and created a new VCE dataset, Kvasir-Capsule [5].…”

Section: Datasetsmentioning

confidence: 99%

“…Additionally, we provide a comma separated value (CSV) file named image-labels.csv that gives the mapping between an image (filename), the label for each image, the corresponding video it originate from, and the video frame number. The Kvasir-Capsule [5] dataset will be an open-source dataset available for others scientists, and will later be grown to include more VCE videos, and more labeled and unlabeled samples. Kvasir-Capsule dataset is available from the Open Science Framework (OSF) accessible via the link https://osf.io/dv2ag/.…”

Section: Kvasir-capsulementioning

confidence: 99%

“…This tf.keras implementation comes with pre-trained weights for both ImageNet and Noisy-Student. We use the weights trained on ImageNet with AutoAugment which are made open source by the authors of EfficientNet on the official github repository 5 .…”

Section: Teacher-student Architecturementioning

confidence: 99%

See 4 more Smart Citations

A self-learning teacher-student framework for gastrointestinal image classification

Gjestang

Hicks

Thambawita

et al. 2021

2021 IEEE 34th International Symposium on Computer-Based Medical Systems (CBMS)

Self Cite

View full text Add to dashboard Cite

Medical data is growing at an estimated 2.5 exabytes per year [1]. However, medical data is often sparse and unavailable for the research community, and qualified medical personnel rarely have time for the tedious labeling work required to prepare the data. New screening methods of the gastrointestinal (GI) tract, like video capsule endoscopy (VCE), can help to reduce patients discomfort and help to increase screening capabilities. One of the main reasons why VCE is not more commonly used by medical experts is the amount of data it produces. A high level of extra work is required by the physicians who, depending on the patient, have to look at more than 50,000 frames per examination. To make VCE more accepted and useful data analysis methods such as machine learning can be very useful.Even if a lot of frames are collected per patient they are most of the time showing normal tissue without any relevant finding. This introduces another problem, namely that it is difficult to train a machine learning based method using this data. Existing models often struggle with the challenge of not having enough data that contains anomalies. This often leads to overfitted and not generalisable models. Our work explores ways to help existing models to overcome this problem by utilising a popular sub-category of machine learning called semi-supervised learning. Semi-supervised learning uses a combination of labeled and unlabeled data which allows us to take advantage of large amounts of unlabeled data.In this thesis, we introduce our proposed semi-supervised teacher-student framework. This framework is built specifically to take advantage of vast amount of unlabeled data and consists of three main steps: (1) train a teacher model with labeled data, (2) use the teacher model to infer pseudo labels with unlabeled data, and (3) train a new and larger student model with a combination of labeled images and inferred pseudo labels. These three steps are repeated several times by treating the student as a teacher to relabel the unlabeled data and consequently training a new student.We demonstrate that our framework can be of use for classifying both, VCE and endoscopic colonoscopy images or videos. We demonstrate that our teacher-student model can significantly increase the performance compared to traditional supervisedlearning-based models. We believe that our framework has potential to be a useful addition to existing medical multimedia systems for automatic disease detection, because new data can be continuously added to improve the models performance while in production.i ii I would like to thank my supervisors Pål Halvorsen and Michael Riegler for all the help and motivation which has kept me going throughout my thesis, and for the opportunity of working on this research topic. I also wish to express my gratitude towards the two PhD student; Steven Hicks and Vajira Thambawita. Thank you for all your help, advice and support, of which I was blessed with during late nights, weekends, holidays and a global pandemic. You are the wisest m...

show abstract

Section: Main Contributionsmentioning

confidence: 99%

Section: Requirementmentioning

confidence: 99%

“…In Section 3.1.2 in the next chapter, we will introduce how we gathered data and created a new VCE dataset, Kvasir-Capsule [5].…”

Section: Datasetsmentioning

confidence: 99%

Section: Kvasir-capsulementioning

confidence: 99%

Section: Teacher-student Architecturementioning

confidence: 99%