Downloads · 30 days
45
29% of all-time downloads
sdoerrich97/true_colon_rtdetr_realcolon_s0
true_colon_rtdetr_realcolon_s0 is a object detection model from sdoerrich97. Use it when you need objects located in an image. It is set up for ultralytics. The card lists the license as agpl-3.0.
Official weights for TRUE-Colon: Exposing a Consistent Transfer Asymmetry in Real-Time Polyp Detection (MICCAI 2026 EndoLINA Workshop). This repo holds the rtdetr detector at seed 0, trained on REAL-Colon: 60 complete…
Downloads · 30 days
45
29% of all-time downloads
All-time downloads
154
Public
Repo size
132 MB
Likes
1
Public
Click a slice to open those files.
.pt132 MB · 100%
From the Hugging Face model README
Official weights for TRUE-Colon: Exposing a Consistent Transfer Asymmetry in Real-Time
Polyp Detection (MICCAI 2026 EndoLINA Workshop). This repo holds the rtdetr
detector at seed 0, trained on REAL-Colon: 60 complete, unedited colonoscopy
procedures rather than curated lesion-centric clips.
TRUE-Colon is an evaluation protocol first. Its finding is that detectors trained on curated benchmarks collapse on full procedures, while procedure-trained models like this one suppress non-polyp content and still transfer back to curated data. The detector itself is an off-the-shelf RT-DETR: what is released here is a baseline trained under the paper's protocol, not a new architecture.
This is one of three seeds. Reporting a single seed of this model in isolation would overstate its precision; the paper aggregates all three.
lesion)true_colon evaluation package (© Sebastian Doerrich, Andreas Franz Schwab). Using these
weights in a product carries the AGPL's source-availability obligations.| Backbone | RT-DETR |
| Input | 640 x 640 |
| Classes | 1 (lesion) |
| Training data | REAL-Colon, patient-level split, 86.47% lesion-free training frames |
| Seed | 0 |
| Batch size | 66 |
| Optimizer | SGD, lr0 0.01, momentum 0.9, weight decay 5e-4 (resolved from optimizer=auto) |
| Schedule | 100 epochs, early stopping patience 10, 3-epoch warm-up |
| Initialization | COCO-pretrained, as shipped by Ultralytics |
| Ultralytics version | 8.3.232 |
This is the configuration this checkpoint was trained with. The paper states the recipe in one sentence and otherwise defers to the frameworks' defaults, so the values it does not state — the optimizer, learning rate, momentum, weight decay and warm-up — come from the training run itself.
On the optimizer. The run recorded
optimizer: auto, which is the framework default the paper keeps. Ultralytics resolves that at runtime: SGD when the iteration budget exceeds 10,000, AdamW otherwise. This run reaches 2,585,900 iterations, so it trained with SGD, atlr0=0.01andmomentum=0.9—autosets both, so the run's ownmomentumentry does not apply.
Research on polyp detection and on evaluation methodology for video-based computer-aided detection. The intended use is reproducing and extending the TRUE-Colon benchmark.
This is not a medical device and must not be used for clinical decision-making. The paper's own conclusion is that performance under realistic full-procedure conditions remains insufficient for reliable clinical deployment. It has not been validated prospectively, cleared by any regulator, or tested against the population, scope hardware or bowel-preparation quality of any particular site.
from ultralytics import YOLO # or RTDETR for the RT-DETR variant
from huggingface_hub import hf_hub_download
weights = hf_hub_download("sdoerrich97/true_colon_rtdetr_realcolon_s0", "model.pt")
model = YOLO(weights)
results = model.predict("frame.jpg", conf=0.30)
Then evaluate under the protocol:
import true_colon as tc
evaluator = tc.create_evaluator(dataset="realcolon", id_source="pred_string")
result = evaluator.evaluate_paths("gt.json", "predictions.json")
Evaluate with the TRUE-Colon protocol rather than COCO mAP alone: the point of the benchmark is that localization accuracy alone does not predict deployment behaviour. The paper reports this architecture at a matched false-alert operating point of tau = 0.30* on REAL-Colon, targeting a frame-level false-positive rate of 4-5%.
For the full result tables, see the paper. Per-seed numbers are not reproduced on this card, because a single seed's values are not the quantity the paper reports.
@article{doerrich2026truecolon,
title={TRUE-Colon: Exposing a Consistent Transfer Asymmetry in Real-Time Polyp Detection},
author={Sebastian Doerrich and Andreas Franz Schwab and Francesco {Di Salvo} and Shyam Nandan Rai and Hanh Huyen My Nguyen and Christian Ledig},
year={2026},
eprint={2608.13711},
archivePrefix={arXiv},
primaryClass={eess.IV},
url={https://arxiv.org/abs/2608.13711},
}
For questions or issues, please open an issue in the GitHub repository, or contact [email protected].