Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces YOLO-PVC, a novel framework that consolidates slice-wise 2D detections into coherent 3D representations for liver tumor localization in MRI. By enforcing depth continuity and utilizing robust percentile statistics for bounding box aggregation, the method effectively addresses the challenges of fragmented predictions and missing detections. Experimental results show that YOLO-PVC achieves a significant improvement in 3D Intersection over Union (IoU) scores, reaching 0.710 with the calibrated variant, indicating its efficacy in clinical applications.
Fragmented 2D detections can be transformed into accurate 3D localizations, achieving a remarkable 0.710 IoU in liver tumor MRI analysis.
Slice-wise 2D object detectors are increasingly applied to volumetric data due to their computational efficiency and scalability, yet they often yield fragmented and unstable predictions along the depth axis. We propose YOLO-PVC, a lightweight and model-agnostic framework for 2D-to-3D consolidation of slice-wise detections. The method enforces depth continuity, aggregates bounding box coordinates using robust percentile statistics, and further refines axial extent through a lightweight MLP-based calibration module. Unlike na\"ive stacking or averaging strategies, YOLO-PVC explicitly addresses missing detections and outlier slices along the depth dimension. Experiments on 3D liver MRI volumes across three tumor categories demonstrate consistent improvements over multiple aggregation baselines. The heuristic PVC achieves an overall $\mathrm{IoU}_{3D}$ of $0.665$, while the calibrated variant further improves performance to $0.710$, with high planar overlap ($\mathrm{BEV\ IoU} \approx 0.78$). These results demonstrate that structured geometric consolidation provides an effective and practical solution for volumetric liver tumor localization in clinical MRI.