检索结果-内蒙古大学图书馆

MRIShift: Disentangled Representation learning for 3D MRI Lesion Segmentation Under Distributional Shifts

学校读者我要写书评

暂无评论

MRIShift: Disentangled Representation Learning for 3D MRI Le...

European Workshop on Visual Information Processing (EUVIP)

作者： Umaima Rahman Guangyi Chen Kun Zhang Computer Vision Department Mohamed Bin Zayed University of Artificial Intelligence Abu Dhabi UAE Machine Learning Department Carnegie Mellon University Pittsburgh PA USA

ISBN: (数字)9798350391145

ISBN: (纸本)9798350391152

In the clinical environment, heterogeneity of data coming from different centers poses challenges where the data may not conform to the assumption of being independent and identically distributed (i.i.d). As a result, it is essential to develop an approach that can generate better predictions even on samples it did not encounter during training. In this paper, we propose an unsupervised domain adaptation approach using disentangled representation learning to enhance the transferability of the 3D segmentation model for medical samples belonging to different distributions unseen during training. When dealing with medical image segmentation, it is essential to understand how features relate to voxel-wise labels. To achieve this, we focus on minimizing changes between data from different sources. This involves organizing the data into two categories: invariant components that remain consistent across different sources, and variant components that exhibit changes. By ensuring that domain shifts primarily affect the changing components, we aim to preserve the identifiability of latent variables. This helps in identifying joint data-label distributions within the target domain. We also introduce a new loss function that integrates image and feature level reconstruction with segmentation loss, enhancing the model's performance. Our experiments on real-world medical data demonstrate improved segmentation accuracy within known distributions and a notable 5% enhancement in handling out-of-distribution samples. Additionally, our method enhances the robustness of the model as demonstrated by the improvement in performance metrics such as R-AUC.

关键词： Training Image segmentation Adaptation models Visualization Three-dimensional displays Accuracy Disentangled representation learning Distributed databases Lesions Biomedical imaging

How to Choose a Reinforcement-learning Algorithm

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Bongratz, Fabian Golkov, Vladimir Mautner, Lukas Libera, Luca Della Heetmeyer, Frederik Czaja, Felix Rodemann, Julian Cremers, Daniel Computer Vision Group Technical University of Munich Germany Munich Center for Machine Learning Germany Department of Statistics Ludwig-Maximilians-Universität Munich Germany

The field of reinforcement learning offers a large variety of concepts and methods to tackle sequential decision-making problems. This variety has become so large that choosing an algorithm for a task at hand can be challenging. In this work, we streamline the process of choosing reinforcement-learning algorithms and action-distribution families. We provide a structured overview of existing methods and their properties, as well as guidelines for when to choose which methods. An interactive version of these guidelines is available online at https://***/.MSC Codes 62M45 © 2024, CC BY-NC-SA.

关键词： Adversarial machine learning

CVPR19 Tracking and Detection Challenge: How crowded can it get?

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Dendorfer, Patrick Rezatofighi, Hamid Milan, Anton Shi, Javen Cremers, Daniel Reid, Ian Roth, Stefan Schindler, Konrad Leal-Taixa, Laura Dynamic Vision and Learning Group at Tum Munich Germany Australian Institute for Machine Learning and School of Computer Science at University of Adelaide. Amazon Berlin Germany Photogrammetry and Remote Sensing Group at Eth Zurich Switzerland Computer Vision Group at Tum Munich Germany Department of Computer Science Technische Universität Darmstadt Germany

Standardized benchmarks are crucial for the majority of computer vision applications. Although leaderboards and ranking tables should not be over-claimed, benchmarks often provide the most objective measure of performance and are therefore important guides for research. The benchmark for Multiple Object Tracking, MOTChallenge, was launched with the goal to establish a standardized evaluation of multiple object tracking methods. The challenge focuses on multiple people tracking, since pedestrians are well studied in the tracking community, and precise tracking and detection has high practical relevance. Since the first release, MOT15 [5], MOT16 [8], and MOT17 [8] have tremendously contributed to the community by introducing a clean dataset and precise framework to benchmark multi-object trackers. In this paper, we present our CVPR19 benchmark, consisting of 8 new sequences depicting very crowded challenging scenes. The benchmark will be presented at the 4th BMTT MOT Challenge Workshop at the computer vision and Pattern Recognition Conference (CVPR) 2019, and will evaluate the state-of-the-art in multiple object tracking whend handling extremely crowded scenarios. Copyright © 2019, The Authors. All rights reserved.

关键词： computer vision

MOT20: A benchmark for multi object tracking in crowded scenes

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Dendorfer, Patrick Rezatofighi, Hamid Milan, Anton Shi, Javen Cremers, Daniel Reid, Ian Roth, Stefan Schindler, Konrad Leal-Taixé, Laura Dynamic Vision and Learning Group at TUM Munich Germany Australian Institute for Machine Learning School of Computer Science University of Adelaide Amazon Berlin Germany Photogrammetry and Remote Sensing Group ETH Zurich Switzerland Computer Vision Group at TUM Munich Germany Department of Computer Science Technische Universität Darmstadt Germany

Standardized benchmarks are crucial for the majority of computer vision applications. Although leaderboards and ranking tables should not be over-claimed, benchmarks often provide the most objective measure of performance and are therefore important guides for research. The benchmark for Multiple Object Tracking, MOTChallenge, was launched with the goal to establish a standardized evaluation of multiple object tracking methods. The challenge focuses on multiple people tracking, since pedestrians are well studied in the tracking community, and precise tracking and detection has high practical relevance. Since the first release, MOT15 [5], MOT16 [8], and MOT17 [8] have tremendously contributed to the community by introducing a clean dataset and precise framework to benchmark multi-object trackers. In this paper, we present our MOT20 benchmark, consisting of 8 new sequences depicting very crowded challenging scenes. The benchmark was presented first at the 4th BMTT MOT Challenge Workshop at the computer vision and Pattern Recognition Conference (CVPR) 2019, and gives to chance to evaluate state-of-the-art methods for multiple object tracking when handling extremely crowded scenarios. Copyright © 2020, The Authors. All rights reserved.

关键词： computer vision

Smoothing splines for discontinuous signals

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Storath, Martin Weinmann, Andreas Lab for Mathematical Methods in Computer Vision and Machine Learning Technische Hochschule Würzburg-Schweinfurt Schweinfurt Germany Department of Mathematics and Natural Sciences Hochschule Darmstadt Darmstadt Germany

Smoothing splines are twice differentiable by construction, so they cannot capture potential discontinuities in the underlying signal. In this work, we consider a special case of the weak rod model of Blake and Zisserman (1987) that allows for discontinuities penalizing their number by a linear term. The corresponding estimates are cubic smoothing splines with discontinuities (CSSD) which serve as representations of piecewise smooth signals and facilitate exploratory data analysis. However, computing the estimates requires solving a non-convex optimization problem. So far, efficient and exact solvers exist only for a discrete approximation based on equidistantly sampled data. In this work, we propose an efficient solver for the continuous minimization problem with non-equidistantly sampled data. Its worst case complexity is quadratic in the number of data points, and if the number of detected discontinuities scales linearly with the signal length, we observe linear growth in runtime. This efficient algorithm allows to use cross validation for automatic selection of the hyperparameters within a reasonable time frame on standard hardware. We provide a reference implementation and supplementary material. We demonstrate the applicability of the approach for the aforementioned tasks using both simulated and real *** Codes 65D07, 65K10, 62G08 Copyright © 2022, The Authors. All rights reserved.

关键词： Convex optimization

Optimising for Interpretability: Convolutional Dynamic Alignment Networks

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Böhle, Moritz Fritz, Mario Schiele, Bernt The Department of Computer Vision and Machine Learning Max Planck Institute for Informatics Saarbrücken66123 Germany The CISPA Helmholtz Center for Information Security Saarbrücken66123 Germany

We introduce a new family of neural network models called Convolutional Dynamic Alignment Networks (CoDA Nets), which are performant classifiers with a high degree of inherent interpretability. Their core building blocks are Dynamic Alignment Units (DAUs), which are optimised to transform their inputs with dynamically computed weight vectors that align with task-relevant patterns. As a result, CoDA Nets model the classification prediction through a series of input-dependent linear transformations, allowing for linear decomposition of the output into individual input contributions. Given the alignment of the DAUs, the resulting contribution maps align with discriminative input patterns. These model-inherent decompositions are of high visual quality and outperform existing attribution methods under quantitative metrics. Further, CoDA Nets constitute performant classifiers, achieving on par results to ResNet and VGG models on e.g. CIFAR-10 and TinyImagenet. Lastly, CoDA Nets can be combined with conventional neural network models to yield powerful classifiers that more easily scale to complex datasets such as Imagenet whilst exhibiting an increased interpretable depth, i.e., the output can be explained well in terms of contributions from intermediate layers within the network. Copyright © 2021, The Authors. All rights reserved.

关键词： Linear transformations

Deep deterministic uncertainty for semantic segmentation

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Mukhoti, Jishnu van Amersfoort, Joost Torr, Philip H.S. Gal, Yarin Oxford Applied & Theoretical Machine Learning Group Department of Computer Science University of Oxford Oxford United Kingdom Torr Vision Group Department of Engineering Science University of Oxford Oxford United Kingdom

We extend Deep Deterministic Uncertainty (DDU) (Mukhoti et al., 2021), a method for uncertainty estimation using feature space densities, to semantic segmentation. DDU enables quantifying and disentangling epistemic and aleatoric uncertainty in a single forward pass through the model. We study the similarity of feature representations of pixels at different locations for the same class and conclude that it is feasible to apply DDU location independently, which leads to a significant reduction in memory consumption compared to pixel dependent DDU. Using the DeepLab-v3+ architecture on Pascal VOC 2012, we show that DDU improves upon MC Dropout and Deep Ensembles while being significantly faster to compute. Copyright © 2021, The Authors. All rights reserved.

关键词： Pixels

B-cos Alignment for Inherently Interpretable CNNs and vision Transformers

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Böhle, Moritz Singh, Navdeeppal Fritz, Mario Schiele, Bernt Department of Computer Vision and Machine Learning Max Planck Institute for Informatics Saarland Informatics Campus Saarbrucken66123 Germany CISPA Helmholtz Center for Information Security Saarbrucken66123 Germany

We present a new direction for increasing the interpretability of deep neural networks (DNNs) by promoting weight-input alignment during training. For this, we propose to replace the linear transformations in DNNs by our novel B-cos transformation. As we show, a sequence (network) of such transformations induces a single linear transformation that faithfully summarises the full model computations. Moreover, the B-cos transformation is designed such that the weights align with relevant signals during optimisation. As a result, those induced linear transformations become highly interpretable and highlight task-relevant features. Importantly, the B-cos transformation is designed to be compatible with existing architectures and we show that it can easily be integrated into virtually all of the latest state of the art models for computer vision-e.g. ResNets, DenseNets, ConvNext models, as well as vision Transformers-by combining the B-cos-based explanations with normalisation and attention layers, all whilst maintaining similar accuracy on ImageNet. Finally, we show that the resulting explanations are of high visual quality and perform well under quantitative interpretability metrics. Copyright © 2023, The Authors. All rights reserved.

关键词： Linear transformations

Underwater Object Detection Enhancement via Channel Stabilization

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Ali, Muhammad Khan, Salman Department of Machine Learning Mohamed bin Zayed University of AI Abu Dhabi United Arab Emirates Department of Computer Vision Mohamed bin Zayed University of AI Abu Dhabi United Arab Emirates

The complex marine environment exacerbates the challenges of object detection manifold. With the advent of the modern era, marine trash presents a danger to the aquatic ecosystem, and it has always been challenging to address this issue with a complete grip. Therefore, there is a significant need to precisely detect marine deposits and locate them accurately in challenging aquatic surroundings. To ensure the safety of the marine environment caused by waste, the deployment of underwater object detection is a crucial tool to mitigate the harm of such waste. Our work explains the image enhancement strategies used and experiments exploring the best detection obtained after applying these methods. Specifically, we evaluate Detectron 2’s backbone performance using different base models and configurations for the underwater detection task. We first propose a novel channel stabilization technique on top of a simplified image enhancement model to help reduce haze and color cast in training images. The proposed procedure shows improved results on multi-scaled objects present in the data set. Following image processing, we investigate various backbones in Detectron2 to provide the best detection accuracy for these images. In addition, we use a sharpening filter with augmentation techniques. This highlights the profile of the object, which helps us easily recognize it. We demonstrate our results by verifying them on the TrashCan Data set, both instance and material versions. We then explore the best-performing backbone method for this setting. We apply our channel stabilization and augmentation methods to the best-performing technique. We also compare our detection results from Detectron2 using the best backbones with those from Deformable Transformer. The detection result for small-sized objects in the instance-version of TrashCan 1.0 gives us a 9.53% absolute increase in average precision, while for the bounding box, we get an absolute gain of 7% compared to the baseline. The

关键词： Underwater imaging