AlignShift: Bridging the Gap of Imaging Thickness in 3D Anisotropic Volumes

Yang, Jiancheng; Yi, He; Huang, Xiaoyang; Xu, Jingwei; Ye, Xiaodan; Tao, Guangyu; Ni, Bingbing

doi:10.1007/978-3-030-59719-1_55

Cited by 22 publications

(29 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As shown in Table 1, our method brings promising detection performance improvements for all baselines. The improvements of Faster R-CNN [27], 9-slice 3DCE, and MVP-Net are more pronounced than those of MULAN w/o SRL [8] and AlignShift [9]. This is because MULAN and AlignShift introduce extra weakly segmentation mask generated from radiologist-annotated RECIST labels.…”

Section: Lesion Detection Performancementioning

confidence: 98%

Bounding Maps for Universal Lesion Detection

Han

Zhou

2020

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Universal Lesion Detection (ULD) in computed tomography plays an essential role in computer-aided diagnosis. Promising ULD results have been reported by coarse-to-fine two-stage detection approaches, but such two-stage ULD methods still suffer from issues like imbalance of positive v.s. negative anchors during object proposal and insufficient supervision problem during localization regression and classification of the region of interest (RoI) proposals. While leveraging pseudo segmentation masks such as bounding map (BM) can reduce the above issues to some degree, it is still an open problem to effectively handle the diverse lesion shapes and sizes in ULD. In this paper we propose a BM-based conditional training for two-stage ULD, which can (i) reduce positive vs. negative anchor imbalance via a BM-based conditioning (BMC) mechanism for anchor sampling instead of traditional IoU-based rule; and (ii) adaptively compute size-adaptive BM (ABM) from lesion boundingbox, which is used for improving lesion localization accuracy via ABMsupervised segmentation. Experiments with four state-of-the-art methods show that the proposed approach can bring an almost free detection accuracy improvement without requiring expensive lesion mask annotations.

show abstract

Section: Lesion Detection Performancementioning

confidence: 98%

Bounding Maps for Universal Lesion Detection

Han

Zhou

2020

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…In this section, we briefly review the 3D context fusion operators that enable 2D pretraining, including (a) no fusion, (b) I3D [2], (c) P3D [13], (d) ACS [23] and (e) Shift [9,22]. As an emerging technique, 3D context fusion operator leverages advantages of both 2D pretraining and 3D context modeling.…”

Section: Preliminary: 3d Context Fusion Operators With 2d Pretrainingmentioning

confidence: 99%

“…For simplicity, only cases with same padding are considered here. Apart from convolutions, we simply convert 2D pooling and normalization into 3D [22]. We then introduce each operator as follows: I3D [2], (c) P3D [13], (d) ACS [23], (e) Shift [9,22] and (f) the proposed A3D.…”

Section: Preliminary: 3d Context Fusion Operators With 2d Pretrainingmentioning

confidence: 99%

“…Recently, there have been a family of techniques that enable building 3D networks with 2D pretraining [2,13,23,9,22], we refer to it as 3D context fusion operators. See Sec.…”

Section: Introductionmentioning

confidence: 99%

“…These operators learn 3D representations while their (partial) learnable weights can be initialized from 2D convolutional kernels. Existing 3D context fusion operators are convolutionlike, i.e., either axial convolutions to fuse slice-wise information [2,13,23] or shifting adjacent slices [9,22]. Therefore, these operators are designed to be spatially symmetric: each 2D slice is operated identically.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Asymmetric 3D Context Fusion for Universal Lesion Detection

Yang,

He,

Kuang

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Modeling 3D context is essential for high-performance 3D medical image analysis. Although 2D networks benefit from large-scale 2D supervised pretraining, it is weak in capturing 3D context. 3D networks are strong in 3D context yet lack supervised pretraining. As an emerging technique, 3D context fusion operator, which enables conversion from 2D pretrained networks, leverages the advantages of both and has achieved great success. Existing 3D context fusion operators are designed to be spatially symmetric, i.e., performing identical operations on each 2D slice like convolutions. However, these operators are not truly equivariant to translation, especially when only a few 3D slices are used as inputs. In this paper, we propose a novel asymmetric 3D context fusion operator (A3D), which uses different weights to fuse 3D context from different 2D slices. Notably, A3D is NOT translation-equivariant while it significantly outperforms existing symmetric context fusion operators without introducing large computational overhead. We validate the effectiveness of the proposed method by extensive experiments on DeepLesion benchmark, a large-scale public dataset for universal lesion detection from computed tomography (CT). The proposed A3D consistently outperforms symmetric context fusion operators by considerable margins, and establishes a new state of the art on DeepLesion. To facilitate open research, our code and model in PyTorch is available at https://github.com/M3DV/AlignShift.

show abstract

Conditional Training with Bounding Map for Universal Lesion Detection

Chen

Han

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

AlignShift: Bridging the Gap of Imaging Thickness in 3D Anisotropic Volumes

Cited by 22 publications

References 27 publications

Bounding Maps for Universal Lesion Detection

Bounding Maps for Universal Lesion Detection

Asymmetric 3D Context Fusion for Universal Lesion Detection

Conditional Training with Bounding Map for Universal Lesion Detection

Contact Info

Product

Resources

About