-
Deep learning architectures for automated image segmentation
Authors:
Debleena Sengupta
Abstract:
Image segmentation is widely used in a variety of computer vision tasks, such as object localization and recognition, boundary detection, and medical imaging. This thesis proposes deep learning architectures to improve automatic object localization and boundary delineation for salient object segmentation in natural images and for 2D medical image segmentation. First, we propose and evaluate a nove…
▽ More
Image segmentation is widely used in a variety of computer vision tasks, such as object localization and recognition, boundary detection, and medical imaging. This thesis proposes deep learning architectures to improve automatic object localization and boundary delineation for salient object segmentation in natural images and for 2D medical image segmentation. First, we propose and evaluate a novel dilated dense encoder-decoder architecture with a custom dilated spatial pyramid pooling block to accurately localize and delineate boundaries for salient object segmentation. The dilation offers better spatial understanding and the dense connectivity preserves features learned at shallower levels of the network for better localization. Tested on three publicly available datasets, our architecture outperforms the state-of-the-art for one and is very competitive on the other two. Second, we propose and evaluate a custom 2D dilated dense UNet architecture for accurate lesion localization and segmentation in medical images. This architecture can be utilized as a stand-alone segmentation framework or used as a rich feature extracting backbone to aid other models in medical image segmentation. Our architecture outperforms all baseline models for accurate lesion localization and segmentation on a new dataset. We furthermore explore the main considerations that should be taken into account for 3D medical image segmentation, among them preprocessing techniques and specialized loss functions.
△ Less
Submitted 19 September, 2019;
originally announced September 2019.
-
Deep Active Lesion Segmentation
Authors:
Ali Hatamizadeh,
Assaf Hoogi,
Debleena Sengupta,
Wuyue Lu,
Brian Wilcox,
Daniel Rubin,
Demetri Terzopoulos
Abstract:
Lesion segmentation is an important problem in computer-assisted diagnosis that remains challenging due to the prevalence of low contrast, irregular boundaries that are unamenable to shape priors. We introduce Deep Active Lesion Segmentation (DALS), a fully automated segmentation framework for that leverages the powerful nonlinear feature extraction abilities of fully Convolutional Neural Networks…
▽ More
Lesion segmentation is an important problem in computer-assisted diagnosis that remains challenging due to the prevalence of low contrast, irregular boundaries that are unamenable to shape priors. We introduce Deep Active Lesion Segmentation (DALS), a fully automated segmentation framework for that leverages the powerful nonlinear feature extraction abilities of fully Convolutional Neural Networks (CNNs) and the precise boundary delineation abilities of Active Contour Models (ACMs). Our DALS framework benefits from an improved level-set ACM formulation with a per-pixel-parameterized energy functional and a novel multiscale encoder-decoder CNN that learns an initialization probability map along with parameter maps for the ACM. We evaluate our lesion segmentation model on a new Multiorgan Lesion Segmentation (MLS) dataset that contains images of various organs, including brain, liver, and lung, across different imaging modalities---MR and CT. Our results demonstrate favorable performance compared to competing methods, especially for small training datasets. Source code : $\text{https://github.com/ahatamiz/dals}$
△ Less
Submitted 30 August, 2020; v1 submitted 19 August, 2019;
originally announced August 2019.
-
Temporo-Spatial Collaborative Filtering for Parameter Estimation in Noisy DCE-MRI Sequences: Application to Breast Cancer Chemotherapy Response
Authors:
Xia Zhu,
Dipanjan Sengupta,
Andrew Beers,
Kalpathy-Cramer Jayashree,
Theodore L. Willke
Abstract:
Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a minimally invasive imaging technique which can be used for characterizing tumor biology and tumor response to radiotherapy. Pharmacokinetic (PK) estimation is widely used for DCE-MRI data analysis to extract quantitative parameters relating to microvascu- lature characteristics of the cancerous tissues. Unavoidable noise corruptio…
▽ More
Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a minimally invasive imaging technique which can be used for characterizing tumor biology and tumor response to radiotherapy. Pharmacokinetic (PK) estimation is widely used for DCE-MRI data analysis to extract quantitative parameters relating to microvascu- lature characteristics of the cancerous tissues. Unavoidable noise corruption during DCE-MRI data acquisition has a large effect on the accuracy of PK estimation. In this paper, we propose a general denoising paradigm called gather- noise attenuation and reduce (GNR) and a novel temporal-spatial collaborative filtering (TSCF) denoising technique for DCE-MRI data. TSCF takes advantage of temporal correlation in DCE-MRI, as well as anatomical spatial similar- ity to collaboratively filter noisy DCE-MRI data. The proposed TSCF denoising algorithm decreases the PK parameter normalized estimation error by 57% and improves the structural similarity of PK parameter estimation by 86% com- pared to baseline without denoising, while being an order of magnitude faster than state-of-the-art denoising methods. TSCF improves the univariate linear regression (ULR) c-statistic value for early prediction of pathologic response up to 18%, and shows complete separation of pathologic complete response (pCR) and non-pCR groups on a challenge dataset.
△ Less
Submitted 2 March, 2018;
originally announced March 2018.
