Establishing ground truth of polyp size, morphology, and volume using three-dimensional scanning
Bibliographic record
Abstract
Establishing ground truth of polyp size, morphology, and volume using three-dimensional scanning Comprehensive artificial intelligence (AI) solutions that should cover detecting polyp morphology, size, and volume are being developed. To train such AI solutions, it is essential to obtain high quality reference (ground truth data) on polyp morphology, size, and volume. We explored the feasibility of three-dimensional (3D) scanning of colorectal polyps to obtain computer-aided design (CAD) 3D rendering information on polyp morphology, size, and volume. During the colonoscopy, a novel virtual scale endoscope (VSE; Scale eye, EW10-VM01; Fujifilm, Tokyo, Japan) with an integrated virtual scale function was used, which allows for polyp size measurement [ 1 ] [ 2 ] [ 3 ] ([ Video 1 ]). Two detected polyps had their size measured during the colonoscopy using VSE ([ Fig. 1 ] a,d ), then were removed from the colon as intact en bloc specimens with a healthy resection margin. The polyps were measured using a digital Vernier caliper (eSync with 32 feeler gauge with 0.01-mm intervals) directly after resection ([ Fig. 1 ] b,e ). The polyps were then 3D scanned for volumetric information (Artec Space Spider; Artec 3D, Luxembourg) [ 4 ] ([ Fig. 1 ] c,f ). The 3D scanner processes up to 1 million individual points per second at 7.5 frames per second with a resolution of 0.1 mm. Polyp 1 was measured as being 8 mm using the VSE, 8.88 mm using the caliper, and 9.68 mm in the 3D model. The total polyp volume was 291.8 mm 3 . Polyp 2 was measured as being 14 mm using the VSE, 15.16 mm using the caliper, and 15.20 mm in the 3D model. The total polyp volume was 968.1 mm 3 . Quality: mobile 360 480 720 Download Two polyps are measured using a virtual scale endoscope (VSE) and then by subsequent 3D rendering after their resection.Video 1 Fig. 1 Images showing the measurement of two polyps using: a,d the virtual scale endoscope; b,e a vernier caliper; c,f 3D rendering after the polyps’ resection. We found that it is possible to obtain spatial polyp information through 3D scanning. 3D scanning of polyps can capture their shape, geometries, and textures, translating them into data files in which size, volume, and shape becomes measurable. These 3D models can then be used in the context of AI development to provide ground truth data to train models to automatically recognize size, volume, and Paris classification. Endoscopy_UCTN_Code_CPL_1AJ_2AB Endoscopy E-Videos https://eref.thieme.de/e-videos E-Videos is an open access online section of the journal Endoscopy , reporting on interesting cases and new techniques in gastroenterological endoscopy. All papers include a high-quality video and are published with a Creative Commons CC-BY license. Endoscopy E-Videos qualify for HINARI discounts and waivers and eligibility is automatically checked during the submission process. We grant 100% waivers to articles whose corresponding authors are based in Group A countries and 50% waivers to those who are based in Group B countries as classified by Research4Life (see: https://www.research4life.org/access/eligibility/ ). This section has its own submission website at https://mc.manuscriptcentral.com/e-videos . Publication History Article published online: 21 December 2023 © 2023. The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. (https://creativecommons.org/licenses/by/4.0/). Georg Thieme Verlag KG Rüdigerstraße 14, 70469 Stuttgart, Germany
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".