Image-Based Breast Implant Detection for Mammography Dataset Curation and Near-Real-Time Deployment: Comparing Foundation Models and Task-Specific Convolutional Models
Abstract
Purpose: To evaluate the performance-feasibility tradeoffs of foundation models (FMs) and task-specific convolutional neural networks (CNNs) trained from scratch for breast implant classification in 2D mammography, with emphasis on suitability for near real-time clinical deployment. Methods: We evaluated four models: two FMs (RAD-DINO and MammoCLIP) and two CNNs trained from scratch for implant prediction (ResNet18 and our lightweight ResNetLite). Using the Emory Breast Imaging Dataset, 5,000 unilateral screening mammograms were used for training/validation and 1,000 manually reviewed unilateral images were held out for testing. For the FMs, global image embeddings from the pretrained encoder were classified using a support vector machine (SVM). The CNNs were trained end-to-end on 2D mammograms, with ResNetLite optimized via grid search over depth and width to balance accuracy and efficiency. Performance was evaluated using AUROC, sensitivity, specificity, accuracy, embedding visualization, and inference-latency. Results: All models demonstrated strong performance on held-out test data (n = 1,000). MammoCLIP achieved the highest AUROC (0.999) with the quickest training time of 493 seconds. RAD-DINO achieved the highest sensitivity (0.980; accuracy 0.989) but had the slowest inference and training times. ResNet18 and MammoCLIP achieved comparable accuracy (0.985). ResNetLite showed no statistically significant difference from ResNet18 (AUROC 0.993; accuracy 0.976) despite using only 1.4% of ResNet18's parameters, and had the fastest inference time. Conclusion: FMs and task-specific CNN models reliably detect breast implants on 2D mammography. Model selection is best guided by deployment context: MammoCLIP for GPU-equipped hospital settings requiring scalable integration, and lightweight CNNs such as ResNetLite for resource-constrained or edge deployments.
Related
- Feature-aware attention convolutional neural network for efficient pulmonary fibrosis classification from chest CT images with FPGA deployment
- iSL-YOLOv11: global–local attention enhanced small lesion detection in mammography images
- Model‐based image restoration targeting higher‐dose digital mammography: Impact on radiologists' detection of clustered microcalcifications
- Ophthalmic AI With Just Ten Examples: Comparing Multimodal Large Language Models, Vision Foundation Models, and Vision-Language Models
- Estimation and detection of rotor mass imbalance of wind turbines using convolutional neural network
BibTeX
@misc{ishwar2026imagebased,
title = {{Image-Based Breast Implant Detection for Mammography Dataset Curation and Near-Real-Time Deployment: Comparing Foundation Models and Task-Specific Convolutional Models}},
author = {Vasisht Ishwar and Hari Trivedi and Young Seok Jeon and Beatrice Brown-Mulry and Frank Li and Rohan Satya Isaac and Mohammadreza Chavoshi and Judy Wawira Gichoya},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.03817},
eprint = {2610.03817},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.03817},
note = {CC0},
}
RIS
TY - UNPB
TI - Image-Based Breast Implant Detection for Mammography Dataset Curation and Near-Real-Time Deployment: Comparing Foundation Models and Task-Specific Convolutional Models
AU - Vasisht Ishwar
AU - Hari Trivedi
AU - Young Seok Jeon
AU - Beatrice Brown-Mulry
AU - Frank Li
AU - Rohan Satya Isaac
AU - Mohammadreza Chavoshi
AU - Judy Wawira Gichoya
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2610.03817
UR - https://arxiv.org/abs/2610.03817
AB - Purpose: To evaluate the performance-feasibility tradeoffs of foundation models (FMs) and task-specific convolutional neural networks (CNNs) trained from scratch for breast implant classification in 2D mammography, with emphasis on suitability for near real-time clinical deployment. Methods: We evaluated four models: two FMs (RAD-DINO and MammoCLIP) and two CNNs trained from scratch for implant prediction (ResNet18 and our lightweight ResNetLite). Using the Emory Breast Imaging Dataset, 5,000 unilateral screening mammograms were used for training/validation and 1,000 manually reviewed unilateral images were held out for testing. For the FMs, global image embeddings from the pretrained encoder were classified using a support vector machine (SVM). The CNNs were trained end-to-end on 2D mammograms, with ResNetLite optimized via grid search over depth and width to balance accuracy and efficiency. Performance was evaluated using AUROC, sensitivity, specificity, accuracy, embedding visualization, and inference-latency. Results: All models demonstrated strong performance on held-out test data (n = 1,000). MammoCLIP achieved the highest AUROC (0.999) with the quickest training time of 493 seconds. RAD-DINO achieved the highest sensitivity (0.980; accuracy 0.989) but had the slowest inference and training times. ResNet18 and MammoCLIP achieved comparable accuracy (0.985). ResNetLite showed no statistically significant difference from ResNet18 (AUROC 0.993; accuracy 0.976) despite using only 1.4% of ResNet18's parameters, and had the fastest inference time. Conclusion: FMs and task-specific CNN models reliably detect breast implants on 2D mammography. Model selection is best guided by deployment context: MammoCLIP for GPU-equipped hospital settings requiring scalable integration, and lightweight CNNs such as ResNetLite for resource-constrained or edge deployments.
ER -