Comprehensive review of publicly available colonoscopic imaging databases for artificial intelligence research: availability, accessibility, and usability

Houwen, Britt B S L; Nass, Karlijn J.; Vleugels, Jasper L A; Fockens, Paul; Hazewinkel, Yark; Dekker, Evelien

doi:10.1016/j.gie.2022.08.043

Cited by 13 publications

(10 citation statements)

References 146 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…of images), the method used to provide image location (Ground Truth), and whether the dataset includes images without polyps and the percentage of images with polyps in case they are included (Without polyps?). Taking into account the characteristics of the different datasets, a high degree of variability can be observed with respect to the number of images included, and whether they contain not-polyp images or not [8,24]. Nevertheless, the trend in recent years is to create increasingly larger datasets with the presence of images without polyps.…”

Section: Public Datasetsmentioning

confidence: 99%

Negative Samples for Improving Object Detection—A Case Study in AI-Assisted Colonoscopy for Polyp Detection

et al. 2023

View full text Add to dashboard Cite

Deep learning object-detection models are being successfully applied to develop computer-aided diagnosis systems for aiding polyp detection during colonoscopies. Here, we evidence the need to include negative samples for both (i) reducing false positives during the polyp-finding phase, by including images with artifacts that may confuse the detection models (e.g., medical instruments, water jets, feces, blood, excessive proximity of the camera to the colon wall, blurred images, etc.) that are usually not included in model development datasets, and (ii) correctly estimating a more realistic performance of the models. By retraining our previously developed YOLOv3-based detection model with a dataset that includes 15% of additional not-polyp images with a variety of artifacts, we were able to generally improve its F1 performance in our internal test datasets (from an average F1 of 0.869 to 0.893), which now include such type of images, as well as in four public datasets that include not-polyp images (from an average F1 of 0.695 to 0.722).

show abstract

Section: Public Datasetsmentioning

confidence: 99%

Negative Samples for Improving Object Detection—A Case Study in AI-Assisted Colonoscopy for Polyp Detection

et al. 2023

View full text Add to dashboard Cite

show abstract

“…However, as many of them were created to validate CADe, the currently available histopathologically annotated databases are insufficient to validate the performance of characterization by CADx models. 3 Kader et al 4 tackled these problems in this issue of Digestive Endoscopy. With a large training dataset, including images of a sessile serrated lesion (SSL), the authors' self-developed convolutional neural network-based CADx model can characterize colorectal polyps, including SSLs, which is not supported by many CADx models in both narrow band imaging (NBI) and NBI-near focus (NF).…”

mentioning

confidence: 99%

“…Henceforth, proper publicly accessible image databases for AI should include not only pooled images and videos, but also their annotated information (metadata) that is properly organized, taking into account the security of personally identifiable information. Recently, the results of a high-quality, comprehensive review, including databases searched from the scientific literature and online database search engines, were reported by Houwen et al 3 This review identified barriers to data accessibility and the inadequate reporting of important metadata as potential limitations to the appropriate use of databases. In addition, similar systematic reviews that aimed to assess the appropriateness of publicly available datasets for digital health research, including the development of machine-learning algorithms in the fields of dermatology and ophthalmology, identified problems with metadata and accessibility.…”

mentioning

confidence: 99%

“…Currently, as the importance of reproducible and widely available benchmarking tests has come to be recognized, publicly accessible databases of colonoscopic images have been established to provide a fair validation of CAD systems. However, as many of them were created to validate CADe, the currently available histopathologically annotated databases are insufficient to validate the performance of characterization by CADx models 3 …”

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Benchmark test for the characterization of colorectal polyps using a computer‐aided diagnosis with a publicly accessible database

Kamba

Sumiyama

2023

Digestive Endoscopy

View full text Add to dashboard Cite

Generation of a Melanoma and Nevus Data Set From Unstandardized Clinical Photographs on the Internet

Cho,

Navarrete-Dechent,

Daneshjou

et al. 2023

JAMA Dermatol

View full text Add to dashboard Cite

ImportanceArtificial intelligence (AI) training for diagnosing dermatologic images requires large amounts of clean data. Dermatologic images have different compositions, and many are inaccessible due to privacy concerns, which hinder the development of AI.ObjectiveTo build a training data set for discriminative and generative AI from unstandardized internet images of melanoma and nevus.Design, Setting, and ParticipantsIn this diagnostic study, a total of 5619 (CAN5600 data set) and 2006 (CAN2000 data set; a manually revised subset of CAN5600) cropped lesion images of either melanoma or nevus were semiautomatically annotated from approximately 500 000 photographs on the internet using convolutional neural networks (CNNs), region-based CNNs, and large mask inpainting. For unsupervised pretraining, 132 673 possible lesions (LESION130k data set) were also created with diversity by collecting images from 18 482 websites in approximately 80 countries. A total of 5000 synthetic images (GAN5000 data set) were generated using the generative adversarial network (StyleGAN2-ADA; training, CAN2000 data set; pretraining, LESION130k data set).Main Outcomes and MeasuresThe area under the receiver operating characteristic curve (AUROC) for determining malignant neoplasms was analyzed. In each test, 1 of the 7 preexisting public data sets (total of 2312 images; including Edinburgh, an SNU subset, Asan test, Waterloo, 7-point criteria evaluation, PAD-UFES-20, and MED-NODE) was used as the test data set. Subsequently, a comparative study was conducted between the performance of the EfficientNet Lite0 CNN on the proposed data set and that trained on the remaining 6 preexisting data sets.ResultsThe EfficientNet Lite0 CNN trained on the annotated or synthetic images achieved higher or equivalent mean (SD) AUROCs to the EfficientNet Lite0 trained using the pathologically confirmed public data sets, including CAN5600 (0.874 [0.042]; P = .02), CAN2000 (0.848 [0.027]; P = .08), and GAN5000 (0.838 [0.040]; P = .31 [Wilcoxon signed rank test]) and the preexisting data sets combined (0.809 [0.063]) by the benefits of increased size of the training data set.Conclusions and RelevanceThe synthetic data set in this diagnostic study was created using various AI technologies from internet images. A neural network trained on the created data set (CAN5600) performed better than the same network trained on preexisting data sets combined. Both the annotated (CAN5600 and LESION130k) and synthetic (GAN5000) data sets could be shared for AI training and consensus between physicians.

show abstract

Comprehensive review of publicly available colonoscopic imaging databases for artificial intelligence research: availability, accessibility, and usability

Cited by 13 publications

References 146 publications

Negative Samples for Improving Object Detection—A Case Study in AI-Assisted Colonoscopy for Polyp Detection

Negative Samples for Improving Object Detection—A Case Study in AI-Assisted Colonoscopy for Polyp Detection

Benchmark test for the characterization of colorectal polyps using a computer‐aided diagnosis with a publicly accessible database

Generation of a Melanoma and Nevus Data Set From Unstandardized Clinical Photographs on the Internet

Contact Info

Product

Resources

About