arXiv subjects

Anirban Mukhopadhyay

Publications and source records attributed to Anirban Mukhopadhyay.

At least 73 records · Page 4Linked to original sources

arXiv2022

Task-agnostic Continual Hippocampus Segmentation for Smooth Population Shifts

Most continual learning methods are validated in settings where task boundaries are clearly defined and task identity information is available during training and testing. We explore how such methods perform in a task-agnostic setting that more closely resembles dynamic clinical environments with gradual population shifts. We propose ODEx, a holistic solution that combines out-of-distribution detection with continual learning techniques. Validation on two scenarios of hippocampus segmentation shows that our proposed method reliably maintains performance on earlier tasks without losing plasticity.

cs.CV↗

arXiv2022

Distance-based detection of out-of-distribution silent failures for Covid-19 lung lesion segmentation

Automatic segmentation of ground glass opacities and consolidations in chest computer tomography (CT) scans can potentially ease the burden of radiologists during times of high resource utilisation. However, deep learning models are not trusted in the clinical routine due to failing silently on out-of-distribution (OOD) data. We propose a lightweight OOD detection method that leverages the Mahalanobis distance in the feature space and seamlessly integrates into state-of-the-art segmentation pipelines. The simple approach can even augment pre-trained models with clinically relevant uncertainty quantification. We validate our method across four chest CT distribution shifts and two magnetic resonance imaging applications, namely segmentation of the hippocampus and the prostate. Our results show that the proposed method effectively detects far- and near-OOD samples across all explored scenarios.

eess.IV↗

arXiv2022

Continual Hippocampus Segmentation with Transformers

In clinical settings, where acquisition conditions and patient populations change over time, continual learning is key for ensuring the safe use of deep neural networks. Yet most existing work focuses on convolutional architectures and image classification. Instead, radiologists prefer to work with segmentation models that outline specific regions-of-interest, for which Transformer-based architectures are gaining traction. The self-attention mechanism of Transformers could potentially mitigate catastrophic forgetting, opening the way for more robust medical image segmentation. In this work, we explore how recently-proposed Transformer mechanisms for semantic segmentation behave in sequential learning scenarios, and analyse how best to adapt continual learning strategies for this setting. Our evaluation on hippocampus segmentation shows that Transformer mechanisms mitigate catastrophic forgetting for medical image segmentation compared to purely convolutional architectures, and demonstrates that regularising ViT modules should be done with caution.

eess.IV↗

arXiv2022

Improving robustness and calibration in ensembles with diversity regularization

Calibration and uncertainty estimation are crucial topics in high-risk environments. We introduce a new diversity regularizer for classification tasks that uses out-of-distribution samples and increases the overall accuracy, calibration and out-of-distribution detection capabilities of ensembles. Following the recent interest in the diversity of ensembles, we systematically evaluate the viability of explicitly regularizing ensemble diversity to improve calibration on in-distribution data as well as under dataset shift. We demonstrate that diversity regularization is highly beneficial in architectures, where weights are partially shared between the individual members and even allows to use fewer ensemble members to reach the same level of robustness. Experiments on CIFAR-10, CIFAR-100, and SVHN show that regularizing diversity can have a significant impact on calibration and robustness, as well as out-of-distribution detection.

cs.LG↗

arXiv2022

Disentanglement enables cross-domain Hippocampus Segmentation

Limited amount of labelled training data are a common problem in medical imaging. This makes it difficult to train a well-generalised model and therefore often leads to failure in unknown domains. Hippocampus segmentation from magnetic resonance imaging (MRI) scans is critical for the diagnosis and treatment of neuropsychatric disorders. Domain differences in contrast or shape can significantly affect segmentation. We address this issue by disentangling a T1-weighted MRI image into its content and domain. This separation enables us to perform a domain transfer and thus convert data from new sources into the training domain. This step thus simplifies the segmentation problem, resulting in higher quality segmentations. We achieve the disentanglement with the proposed novel methodology 'Content Domain Disentanglement GAN', and we propose to retrain the UNet on the transformed outputs to deal with GAN-specific artefacts. With these changes, we are able to improve performance on unseen domains by 6-13% and outperform state-of-the-art domain transfer methods.

eess.IV↗

arXiv2022

On the explicit Galois group of $\mathbb{Q}(\sqrt{a_{1}}, \sqrt{a_{2}}, \dots, \sqrt{a_{n}}, ζ_{d})$ over $\mathbb{Q}$

Let $S= \{ a_{1}, a_{2}, \dots, a_{n} \}$ be a finite set of non-zero integers. In \cite{KBAM21}, Karthick Babu and Anirban Mukhopadhyay calculated the explicit structure of the Galois group of multi-quadratic field $\mathbb{Q}(\sqrt{a_{1}}, \sqrt{a_{2}}, \dots, \sqrt{a_{n}})$ over $\mathbb{Q}$. For a positive integer $d \geq 3$, $ζ_{d}$ denotes the primitive $d$-th root of unity. In this paper, we calculate the explicit structure of the Galois group of $\mathbb{Q}(\sqrt{a_{1}}, \sqrt{a_{2}}, \dots, \sqrt{a_{n}}, ζ_{d})$ over $\mathbb{Q}$ in terms of its action on $ζ_{d}$ and $\sqrt{a_{i}}$ for $1 \leq i \leq n$.

math.NT↗

arXiv2021

Quality monitoring of federated Covid-19 lesion segmentation

Federated Learning is the most promising way to train robust Deep Learning models for the segmentation of Covid-19-related findings in chest CTs. By learning in a decentralized fashion, heterogeneous data can be leveraged from a variety of sources and acquisition protocols whilst ensuring patient privacy. It is, however, crucial to continuously monitor the performance of the model. Yet when it comes to the segmentation of diffuse lung lesions, a quick visual inspection is not enough to assess the quality, and thorough monitoring of all network outputs by expert radiologists is not feasible. In this work, we present an array of lightweight metrics that can be calculated locally in each hospital and then aggregated for central monitoring of a federated system. Our linear model detects over 70% of low-quality segmentations on an out-of-distribution dataset and thus reliably signals a decline in model performance.

eess.IV↗

arXiv2021

How Reliable Are Out-of-Distribution Generalization Methods for Medical Image Segmentation?

The recent achievements of Deep Learning rely on the test data being similar in distribution to the training data. In an ideal case, Deep Learning models would achieve Out-of-Distribution (OoD) Generalization, i.e. reliably make predictions on out-of-distribution data. Yet in practice, models usually fail to generalize well when facing a shift in distribution. Several methods were thereby designed to improve the robustness of the features learned by a model through Regularization- or Domain-Prediction-based schemes. Segmenting medical images such as MRIs of the hippocampus is essential for the diagnosis and treatment of neuropsychiatric disorders. But these brain images often suffer from distribution shift due to the patient's age and various pathologies affecting the shape of the organ. In this work, we evaluate OoD Generalization solutions for the problem of hippocampus segmentation in MR data using both fully- and semi-supervised training. We find that no method performs reliably in all experiments. Only the V-REx loss stands out as it remains easy to tune, while it outperforms a standard U-Net in most cases.

eess.IV↗

arXiv2021

Adversarial Continual Learning for Multi-Domain Hippocampal Segmentation

Deep learning for medical imaging suffers from temporal and privacy-related restrictions on data availability. To still obtain viable models, continual learning aims to train in sequential order, as and when data is available. The main challenge that continual learning methods face is to prevent catastrophic forgetting, i.e., a decrease in performance on the data encountered earlier. This issue makes continuous training of segmentation models for medical applications extremely difficult. Yet, often, data from at least two different domains is available which we can exploit to train the model in a way that it disregards domain-specific information. We propose an architecture that leverages the simultaneous availability of two or more datasets to learn a disentanglement between the content and domain in an adversarial fashion. The domain-invariant content representation then lays the base for continual semantic segmentation. Our approach takes inspiration from domain adaptation and combines it with continual learning for hippocampal segmentation in brain MRI. We showcase that our method reduces catastrophic forgetting and outperforms state-of-the-art continual learning methods.

eess.IV↗

arXiv2021

Detecting when pre-trained nnU-Net models fail silently for Covid-19 lung lesion segmentation

Automatic segmentation of lung lesions in computer tomography has the potential to ease the burden of clinicians during the Covid-19 pandemic. Yet predictive deep learning models are not trusted in the clinical routine due to failing silently in out-of-distribution (OOD) data. We propose a lightweight OOD detection method that exploits the Mahalanobis distance in the feature space. The proposed approach can be seamlessly integrated into state-of-the-art segmentation pipelines without requiring changes in model architecture or training procedure, and can therefore be used to assess the suitability of pre-trained models to new data. We validate our method with a patch-based nnU-Net architecture trained with a multi-institutional dataset and find that it effectively detects samples that the model segments incorrectly.

eess.IV↗

arXiv2021

Simulation-to-Real domain adaptation with teacher-student learning for endoscopic instrument segmentation

Purpose: Segmentation of surgical instruments in endoscopic videos is essential for automated surgical scene understanding and process modeling. However, relying on fully supervised deep learning for this task is challenging because manual annotation occupies valuable time of the clinical experts. Methods: We introduce a teacher-student learning approach that learns jointly from annotated simulation data and unlabeled real data to tackle the erroneous learning problem of the current consistency-based unsupervised domain adaptation framework. Results: Empirical results on three datasets highlight the effectiveness of the proposed framework over current approaches for the endoscopic instrument segmentation task. Additionally, we provide analysis of major factors affecting the performance on all datasets to highlight the strengths and failure modes of our approach. Conclusion: We show that our proposed approach can successfully exploit the unlabeled real endoscopic video frames and improve generalization performance over pure simulation-based training and the previous state-of-the-art. This takes us one step closer to effective segmentation of surgical tools in the annotation scarce setting.

cs.CV↗

arXiv2021

A survey on shape-constraint deep learning for medical image segmentation

Since the advent of U-Net, fully convolutional deep neural networks and its many variants have completely changed the modern landscape of deep learning based medical image segmentation. However, the over dependence of these methods on pixel level classification and regression has been identified early on as a problem. Especially when trained on medical databases with sparse available annotation, these methods are prone to generate segmentation artifacts such as fragmented structures, topological inconsistencies and islands of pixel. These artefacts are especially problematic in medical imaging since segmentation is almost always a pre-processing step for some downstream evaluation. The range of possible downstream evaluations is rather big, for example surgical planning, visualization, shape analysis, prognosis, treatment planning etc. However, one common thread across all these downstream tasks is the demand of anatomical consistency. To ensure the segmentation result is anatomically consistent, approaches based on Markov/ Conditional Random Fields, Statistical Shape Models are becoming increasingly popular over the past 5 years. In this review paper, a broad overview of recent literature on bringing anatomical constraints for medical image segmentation is given, the shortcomings and opportunities of the proposed methods are thoroughly discussed and potential future work is elaborated. We review the most relevant papers published until the submission date. For quick access, important details such as the underlying method, datasets and performance are tabulated.

eess.IV↗

arXiv2021

CycleGAN for Interpretable Online EMT Compensation

Purpose: Electromagnetic Tracking (EMT) can partially replace X-ray guidance in minimally invasive procedures, reducing radiation in the OR. However, in this hybrid setting, EMT is disturbed by metallic distortion caused by the X-ray device. We plan to make hybrid navigation clinical reality to reduce radiation exposure for patients and surgeons, by compensating EMT error. Methods: Our online compensation strategy exploits cycle-consistent generative adversarial neural networks (CycleGAN). 3D positions are translated from various bedside environments to their bench equivalents. Domain-translated points are fine-tuned to reduce error in the bench domain. We evaluate our compensation approach in a phantom experiment. Results: Since the domain-translation approach maps distorted points to their lab equivalents, predictions are consistent among different C-arm environments. Error is successfully reduced in all evaluation environments. Our qualitative phantom experiment demonstrates that our approach generalizes well to an unseen C-arm environment. Conclusion: Adversarial, cycle-consistent training is an explicable, consistent and thus interpretable approach for online error compensation. Qualitative assessment of EMT error compensation gives a glimpse to the potential of our method for rotational error compensation.

cs.CV↗

arXiv2020

Super-Selfish: Self-Supervised Learning on Images with PyTorch

Super-Selfish is an easy to use PyTorch framework for image-based self-supervised learning. Features can be learned with 13 algorithms that span from simple classification to more complex state of theart contrastive pretext tasks. The framework is easy to use and allows for pretraining any PyTorch neural network with only two lines of code. Simultaneously, full flexibility is maintained through modular design choices. The code can be found at https://github.com/MECLabTUDA/Super_Selfish and installed using pip install super-selfish.

cs.CV↗

arXiv2020

Understanding Interpretability by generalized distillation in Supervised Classification

The ability to interpret decisions taken by Machine Learning (ML) models is fundamental to encourage trust and reliability in different practical applications. Recent interpretation strategies focus on human understanding of the underlying decision mechanisms of the complex ML models. However, these strategies are restricted by the subjective biases of humans. To dissociate from such human biases, we propose an interpretation-by-distillation formulation that is defined relative to other ML models. We generalize the distillation technique for quantifying interpretability, using an information-theoretic perspective, removing the role of ground-truth from the definition of interpretability. Our work defines the entropy of supervised classification models, providing bounds on the entropy of Piece-Wise Linear Neural Networks (PWLNs), along with the first theoretical bounds on the interpretability of PWLNs. We evaluate our proposed framework on the MNIST, Fashion-MNIST and Stanford40 datasets and demonstrate the applicability of the proposed theoretical framework in different supervised classification scenarios.

cs.LG↗

arXiv2020

EXCLUVIS: A MATLAB GUI Software for Comparative Study of Clustering and Visualization of Gene Expression Data

Clustering is a popular data mining technique that aims to partition an input space into multiple homogeneous regions. There exist several clustering algorithms in the literature. The performance of a clustering algorithm depends on its input parameters which can substantially affect the behavior of the algorithm. Cluster validity indices determine the partitioning that best fits the underlying data. In bioinformatics, microarray gene expression technology has made it possible to measure the gene expression levels of thousands of genes simultaneously. Many genomic studies, which aim to analyze the functions of some genes, highly rely on some clustering technique for grouping similarly expressed genes in one cluster or partitioning tissue samples based on similar expression values of genes. In this work, an application package called EXCLUVIS (gene EXpression data CLUstering and VISualization) has been developed using MATLAB Graphical User Interface (GUI) environment for analyzing the performances of different clustering algorithms on gene expression datasets. In this application package, the user needs to select a number of parameters such as internal validity indices, external validity indices and number of clusters from the active windows for evaluating the performance of the clustering algorithms. EXCLUVIS compares the performances of K-means, fuzzy C-means, hierarchical clustering and multiobjective evolutionary clustering algorithms. Heatmap and cluster profile plots are used for visualizing the results. EXCLUVIS allows the users to easily find the goodness of clustering solutions as well as provides visual representations of the clustering outcomes.

cs.HC↗

arXiv2020

Bounded gaps between product of two primes in imaginary quadratic number fields

We study the gaps between products of two primes in imaginary quadratic number fields using a combination of the methods of Goldston-Graham-Pintz-Yildirim \cite{GGPY}, and Maynard \cite{MAY}. An important consequence of our main theorem is existence of infinitely many pairs $α_1, α_2$ which are product of two primes in the imaginary quadratic field $K$ such that $|σ(α_1-α_2)|\leq 2$ for all embedding $σ$ of $K$ if the class number of $K$ is one and $|σ(α_1-α_2)|\leq 8$ for all embedding $σ$ of $K$ if the class number of $K$ is two.

math.NT↗

arXiv2020

Endo-Sim2Real: Consistency learning-based domain adaptation for instrument segmentation

Surgical tool segmentation in endoscopic videos is an important component of computer assisted interventions systems. Recent success of image-based solutions using fully-supervised deep learning approaches can be attributed to the collection of big labeled datasets. However, the annotation of a big dataset of real videos can be prohibitively expensive and time consuming. Computer simulations could alleviate the manual labeling problem, however, models trained on simulated data do not generalize to real data. This work proposes a consistency-based framework for joint learning of simulated and real (unlabeled) endoscopic data to bridge this performance generalization issue. Empirical results on two data sets (15 videos of the Cholec80 and EndoVis'15 dataset) highlight the effectiveness of the proposed \emph{Endo-Sim2Real} method for instrument segmentation. We compare the segmentation of the proposed approach with state-of-the-art solutions and show that our method improves segmentation both in terms of quality and quantity.

cs.CV↗