This research paper introduces a weakly supervised deep learning model that achieves comparable performance to fully supervised models in detecting clinically significant prostate cancer (csPCa) from multiparametric MRI, using significantly fewer annotations and demonstrating robustness to domain shifts.
This research paper introduces a novel synthetic vascular model (VaMos) for generating realistic 3D intracranial aneurysm images, demonstrating its efficacy in augmenting training data for a deep learning-based aneurysm detection system and improving detection sensitivity.
本稿では、深層学習を用いた頭蓋内動脈瘤の検出において、合成血管モデルを用いたデータ拡張が有効であることを示しています。
Synthetic data augmentation, using a novel vascular model (VaMos) mimicking cerebral arteries and aneurysms, significantly improves the sensitivity of intracranial aneurysm detection in 3D Convolutional Neural Networks.
本稿では、深層学習を用いてk空間における頭部の動きを推定し、モデルベースの動き補正を用いて画像を復元する新しい手法「SISMIK」を提案する。この手法は、動きがない参照画像を必要とせず、k空間の広範囲な空間周波数領域で学習できる点が特徴である。
This paper introduces SISMIK, a novel deep learning-based method for estimating and correcting in-plane rigid-body motion artifacts in brain MRI by analyzing k-space data, offering a promising solution for improving image quality without relying on motion-free references or introducing hallucinations.
本稿では、スパースビューCBCT再構成における課題に対処するため、組織誘導ニューラルトモグラフィー(TNT)と呼ばれる新しい手法を提案する。TNTは、軟組織と硬組織の強度差に着目し、強度場を4つの構成要素(軟組織と硬組織の形状とテクスチャ)に分解することで、ニューラルフィールドの学習プロセスを簡簡素化し、再構成の精度と効率を向上させる。
By decomposing the intensity field into distinct tissue components and using tissue-guided supervision during training, TNT significantly improves the quality and efficiency of sparse-view CBCT reconstruction.
This research paper introduces a novel method for accelerating multi-spectral MRI scans near metal implants using variable resolution sampling and deep learning-based image reconstruction, demonstrating its potential to reduce scan times while maintaining high image quality.
FANCL, a novel deep learning model, leverages feature-guided attention and curriculum learning to overcome limitations of traditional CNNs in segmenting brain metastases (BMs) from MRI, particularly small and irregular lesions, achieving superior performance compared to existing methods.