Accurate detection of mitotic figures in breast histopathology images is central to tumor grading and prognostic assessment. Precise bounding-box annotation remains labor-intensive and variable because mitotic figures are small, morphologically diverse, and boundary-ambiguous. Point-level supervision reduces annotation cost but lacks scale information, making reliable pseudo-box generation essential. Existing point-supervised pipelines often use static or heuristic thresholds that may become unstable as teacher predictions and proposal-score distributions evolve. We propose DDFMitos-Net, a point-supervised teacher-student framework for distribution-aware proposal filtering. The framework learns initial scale priors from point-guided simulated masks, refines teacher-generated pseudo-boxes through Adaptive Multiple Instance Learning, and uses Distribution-based Dynamic Filtering to integrate classification confidence with point-centered spatial information. Adaptive thresholds are estimated with a truncated Dirichlet Process Mixture Model. Transformation-Scale Learning improves geometric consistency, and Center-Aware Domain Adaptation provides auxiliary scanner-aware feature alignment. On MITOS12, MITOS14, TUPAC16, and MIDOG21, DDFMitos-Net achieved repeated-run F1 scores of 0.837 ± 0.004, 0.716 ± 0.006, 0.785 ± 0.005, and 0.806 ± 0.004, respectively. These results indicate stable and competitive point-supervised mitosis detection using low-cost point-level annotations.
Chang Shu, Qiling Tang, Jian-Chi Yue et al.· IEEE journal of biomedical a...· 0 citations
BACKGROUND AND PURPOSE
The accurate preoperative classification of parotid tumors is crucial in selecting the appropriate treatment options. This study aimed to develop a machine learning model based on clinical, imaging, and laboratory test data to efficiently determine the preoperative nature of primary operable parotid tumors.
METHODS
A total of 779 patients with primary operable parotid tumors receiving surgical treatment from four independent clinical centers were enrolled. Standardized clinical, laboratory and multi-modality imaging indicators were extracted for model construction. We compared diagnostic performance from experienced clinicians' empirical diagnosis and ten machine learning algorithms via receiver operating characteristic (ROC) analysis on internal training/test cohorts and an independent external validation dataset.
RESULT
In total 779 enrolled subjects, empirical clinical prediction yielded an AUC(Area Under the Receiver Operating Characteristic Curve) of 0.83 (95% CI, 0.79-0.87), while the CatBoost model based on integrated multi-source data achieved an AUC of 0.91 (95% CI, 0.85-0.97). The CatBoost model demonstrated a sensitivity of 76%, specificity of 95%, Youden's index of 0.71, and accuracy of 92% for identifying malignant tumors.
CONCLUSION
The results of this study demonstrate that the CatBoost model constructed based on clinical, imaging and laboratory test features exhibits certain auxiliary predictive value in the preoperative assessment of primary operable parotid tumor nature, with its overall performance slightly superior to that of traditional clinical evaluation and logistic regression models.
Danyi Du, Silin Zhang, Chun-Lei Yang et al.· BMC Oral Health· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
