Progress in colonoscopy polyp segmentation is routinely reported through leaderboard comparisons on a small set of public benchmarks. We argue that this apparent progress is difficult to verify: a systematic audit of \textbf{27 papers} published between 2015 and 2026 reveals three structural problems in how the community evaluates models.
arXiv:2609.15130v1 Announce Type: cross Abstract: woma is a real-time foundation model for gastrointestinal endoscopy: a network trained without labels on about a million endoscopy frames, from which...
arXiv:2608.28716v1 Announce Type: cross Abstract: Deep learning segmentation models are often evaluated using geometric metrics such as Dice, HD95, and ASD, yet it remains unclear to what extent impr...
arXiv:2609.10495v1 Announce Type: new Abstract: In real-time colonoscopy, ground-truth annotations are unavailable at inference, so polyp segmentation models can fail silently. We propose Referee-Bas...
arXiv:2608. 13711v1 Announce Type: cross Abstract: Computer-aided detection (CADe) systems for colonoscopy promise to reduce clinical miss rates, yet reliable real-world deployment remains elusive.
SegCol is a new dataset and benchmark for semantic segmentation of colon fold edges and surgical instruments in colonoscopy images, derived from the EndoMapper dataset. It offers manually annotated pixel‑level masks for three instrument classes and thin fold‑edge structures across temporally consistent image sequences, and serves as the basis for the SegCol Challenge within the EndoVis Challenge at MICCAI 2024. The study evaluates supervised segmentation and annotation‑efficient active learning, analyzes various segmentation metrics under structural perturbations, and highlights how metric behavior depends on target structure, underscoring the need for carefully selected evaluation protocols in endoscopic segmentation.
The paper presents a segmentation pipeline for brain metastases in both pre‑ and post‑treatment cases using a 5‑fold nnU‑Net ResEnc‑L ensemble trained on 1,296 four‑modality cases. A rule‑based post‑processing cascade improves the lesion‑wise Dice similarity coefficient (LW‑DSC) for enhancing tumour, tumour core, whole tumour, and resection cavity sub‑regions, achieving LW‑DSC scores of 0.733, 0.751, 0.713, and 0.549 respectively on the official validation leaderboard. The authors conduct a five‑fold out‑of‑fold analysis to validate the robustness of each post‑processing stage, provide a mechanistic explanation of LW‑DSC behaviour, and report thirteen negative results that challenge common intuitions, with all code released under Apache‑2.0.
The paper presents a pragmatic segmentation pipeline for brain metastases in the BraTS 2026 Task 1, using a 5‑fold nnU-Net ResEnc‑L ensemble trained for 1,000 epochs on 1,296 four‑modality cases. A rule‑based post‑processing cascade tuned for the lesion‑wise Dice similarity coefficient (LW‑DSC) improves performance, achieving LW‑DSC scores of 0.733, 0.751, 0.713, and 0.549 on enhancing tumour, tumour core, whole tumour, and resection cavity, respectively. The authors audit each post‑processing stage with a five‑fold out‑of‑fold analysis, confirm two stages as robust, and provide a mechanistic analysis of LW‑DSC, along with thirteen negative results that challenge common intuitions.
The paper presents a four‑layer hierarchical pipeline that constructs a lesion‑centered spatial record from colonoscopy videos without full‑colon 3D reconstruction. It combines a global topological map, lesion‑level spatio‑temporal tracks, on‑demand local 3D reconstruction, and persistent lesion identity across repeated observations, and evaluates the system on four public videos. The results show successful detection of revisit events, accurate lesion identity merging, and superior geometry accuracy compared to a general‑purpose foundation model.
arXiv:2609.01470v1 Announce Type: new Abstract: As AI systems are increasingly used to draft radiology reports, reliably evaluating their clinical quality remains a critical challenge. Large language...
arXiv:2608. 07299v1 Announce Type: cross Abstract: Radiology reports describe clinical observations but do not specify executable segmentation targets.
The paper investigates fiber bundle segmentation in macaque tracer histology, comparing traditional BCE‑Dice loss with topology‑aware losses such as clDice, Betti matching, and Topograph using a frozen DINOv3 backbone. While BCE‑Dice yields the highest Dice score, Topograph achieves comparable Dice with lower topological error and fewer false positives. The authors also introduce Excess32, a spatial diagnostic that reveals oversegmentation issues not captured by conventional detection metrics, demonstrating that detection metrics alone are insufficient for evaluating segmentation quality.
The paper introduces BiCC, a bidirectional connected-component loss that pairs annotation- and prediction-derived partitions to score predicted components on their own scale. By deriving instances from predictions, BiCC directly penalizes false-positive components regardless of size, allowing a balance parameter to control the lesion-wise precision–recall trade-off. Across five datasets, BiCC outperforms existing instance-aware losses such as CC-DiceCE and blob loss in lesion-wise F1, and improves over DiceCE on multiple datasets.