This repository contains the code and configurations for a study submitted as a scientific research article.

This repository contains the code and configurations used in the study titled: "Colonic Polyp Detection with Object Detection Models" by Raluca Portase, and Eugen-Richard Ardelean, published in Computers, 2026.

This project benchmarks modern object detection models (YOLOv8–v12, YOLO26, YOLOE, YOLO-World, RT-DETR) for automatic polyp detection. It evaluates their ability to detect colonic polyps across three publicly available datasets:

• CVC-ClinicDB,
• CVC-ColonDB,
• ETIS-LaribPolypDB,

Our experiments identify YOLO-World and YOLOv11 as the most effective models, offering an excellent trade-off between detection accuracy and computational efficiency.

This project requires a .yaml file to be created for the models to run on each dataset, as shown in this example:

train:  ..\..\data\CVC-ClinicDB\images\train
val:    ..\..\data\CVC-ClinicDB\images\val
test:   ..\..\data\CVC-ClinicDB\images\test
nc: 1
names: [Polyp]

• Name: CVC-ClinicDB
• Description: public colonoscopy polyp segmentation dataset released by the CVC (Computer Vision Center) group in collaboration with Hospital Clínic (Barcelona). It contains 612 frames (with a size of 348 × 288 pixels) extracted from colonoscopy videos (each paired with a pixel-wise polyp mask); it comprises 31 different polyps from 31 sequences from 23 patients.
• Access: Available at 🔗 CVC-ClinicDB Dataset Page

• Name: CVC-ColonDB
• Description: public colonoscopy dataset of still frames containing annotated polyps. It contains 300 polyp frames (with a size of 574 × 500 pixels) selected from colonoscopy videos (drawn from from 13 polyp video sequences from 13 patients) to maximize viewpoint variability per polyp and provided with ground-truth polyp masks.
• Access: Available at 🔗 CVC-ColonDB Dataset Page

• Name: ETIS-LaribPolypDB
• Description: used in the MICCAI 2015 Endoscopic Vision Challenge as the challenge test set. It was assembled by ETIS Lab and Lariboisière Hospital and contains 196 high-definition frames (each with a segmentation mask), collected from dozens of video sequences (comprises 44 different polyps from 34 sequences) and representing a diverse, challenging set of polyps from multiple patients, devices and larger image resolutions.
• Access: Available at 🔗 ETIS-LaribPolypDB Dataset Page

• YOLOv8 to YOLOv12
• YOLO26
• YOLOE: Prompt-guided detection
• YOLO-World: Open-vocabulary object detection
• RT-DETR: Transformer-based object detection

We show here the cross-dataset evaluation for models trained on either CVC-ClinicDB or CVC-ColonDB, then tested on ETIS-LaribPolypDB

If you use this code or reference the models/datasets in your work, please cite:

@article{10.3390/computers15040258,
    author = {Portase, Raluca and Ardelean, Eugen-Richard},
    title = {Colonic Polyp Detection with Object Detection Models},
    journal = {Computers},
    volume = {15},
    year = {2026},
    number = {4},
    article-number = {258},
    url = {https://www.mdpi.com/2073-431X/15/4/258},
    ISSN = {2073-431X},
    DOI = {10.3390/computers15040258}
}

For questions, please contact: 📧 [email protected]