| 1. |
- Andersson, Matilda, et al.
(author)
-
Augmentation Strategies for Self-Supervised Representation Learning from Electrocardiograms
- 2023
-
In: 31st European Signal Processing Conference, EUSIPCO 2023 - Proceedings. - 2219-5491. - 9789464593600 ; , s. 1075-1079
-
Conference paper (peer-reviewed)abstract
- In this paper, we investigate the effects of different augmentation strategies in self-supervised representation learning from electrocardiograms. Our study examines the impact of random resized crop and time out on downstream performance. We also consider the importance of the signal length. Furthermore, instead of using two augmented copies of the sample as a positive pair, we suggest augmenting only one. The second signal is kept as the original signal. These different augmentation strategies are investigated in the context of pre-training and fine-tuning, following the different self-supervised learning frameworks BYOL, SimCLR, and VICReg. We formulate the downstream task as a multi-label classification task using a public dataset containing ECG recordings and annotations. In our experiments, we demonstrate that self-supervised learning can consistently outperform classical supervised learning when configured correctly. These findings are of particular importance in the medical domain, as the medical labeling process is particularly expensive, and clinical ground truth is often difficult to define. We are hopeful that our findings will be a catalyst for further research into augmentation strategies in self-supervised learning to improve performance in the detection of cardiovascular disease.
|
|
| 2. |
- Arvidsson, Ida, et al.
(author)
-
Detection of left bundle branch block and obstructive coronary artery disease from myocardial perfusion scintigraphy using deep neural networks
- 2021
-
In: Medical Imaging 2021 : Computer-Aided Diagnosis - Computer-Aided Diagnosis. - : SPIE. - 1605-7422. - 9781510640238 ; 11597
-
Conference paper (peer-reviewed)abstract
- Myocardial perfusion scintigraphy, which is a non-invasive imaging technique, is one of the most common cardiological examinations performed today, and is used for diagnosis of coronary artery disease. Currently the analysis is performed visually by physicians, but this is both a very time consuming and a subjective approach. These are two of the motivations for why an automatic tool to support the decisions would be useful. We have developed a deep neural network which predicts the occurrence of obstructive coronary artery disease in each of the three major arteries as well as left bundle branch block. Since multiple, or none, of these could have a defect, this is treated as a multi-label classification problem. Due to the highly imbalanced labels, the training loss is weighted accordingly. The prediction is based on two polar maps, captured during stress in upright and supine position, together with additional information such as BMI and angina symptoms. The polar maps are constructed from myocardial perfusion scintigraphy examinations conducted in a dedicated Cadmium-Zinc-Telluride cardio camera (D-SPECT Spectrum Dynamics). The study includes data from 759 patients. Using 5-fold cross-validation we achieve an area under the receiver operating characteristics curve of 0.89 as average on per-vessel level for the three major arteries, 0.94 on per-patient level and 0.82 for left bundle branch block.
|
|
| 3. |
- Arvidsson, Ida, et al.
(author)
-
Domain-adversarial neural network for improved generalization performance of gleason grade classification
- 2020
-
In: Medical Imaging 2020 : Digital Pathology - Digital Pathology. - : SPIE. - 1605-7422. - 9781510634077 ; 11320
-
Conference paper (peer-reviewed)abstract
- When training a deep learning model, the dataset used is of great importance to make sure that the model learns relevant features of the data and that it will be able to generalize to new data. However, it is typically difficult to produce a dataset without some bias toward any specific feature. Deep learning models used in histopathology have a tendency to overfit to the stain appearance of the training data - if the model is trained on data from one lab only, it will usually not be able to generalize to data from other labs. The standard technique to overcome this problem is to use color augmentation of the training data which, artificially, generates more variations for the network to learn. In this work we instead test the use of a so called domain-adversarial neural network, which is designed to prevent the model from being biased towards features that in reality are irrelevant such as the origin of an image. To test the technique, four datasets from different hospitals for Gleason grading of prostate cancer are used. We achieve state of the art results for these particular datasets, and furthermore for two of our three test datasets the approach outperforms the use of color augmentation.
|
|
| 4. |
- Arvidsson, Ida, et al.
(author)
-
Prediction of Obstructive Coronary Artery Disease from Myocardial Perfusion Scintigraphy using Deep Neural Networks
- 2021
-
In: 2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR). - : IEEE COMPUTER SOC. - 1051-4651. - 9781728188089 ; , s. 4442-4449
-
Conference paper (peer-reviewed)abstract
- For diagnosis and risk assessment in patients with stable ischemic heart disease, myocardial perfusion scintigraphy is one of the most common cardiological examinations performed today. There are however many motivations for why an artificial intelligence algorithm would provide useful input to this task. For example to reduce the subjectiveness and save time for the nuclear medicine physicians working with this time consuming task. In this work we have developed a deep learning algorithm for multi-label classification based on a convolutional neural network to estimate the probability of obstructive coronary artery disease in the left anterior artery, left circumflex artery and right coronary artery. The prediction is based on data from myocardial perfusion scintigraphy studies conducted in a dedicated Cadmium-Zinc-Telluride cardio camera (D-SPECT Spectrum Dynamics). Data from 588 patients was available, with stress images in both upright and supine position, as well as a number of auxiliary parameters such as angina symptoms and age. The data was used to train and evaluate the algorithm using 5-fold cross-validation. We achieve state-of-the-art results for this task with an area under the receiver operating characteristics curve of 0.89 as average on per-vessel level and 0.95 on per-patient level.
|
|
| 5. |
- Aström, Kalle, et al.
(author)
-
Extension of Time-Difference-of-Arrival Self Calibration Solutions Using Robust Multilateration
- 2021
-
In: 29th European Signal Processing Conference, EUSIPCO 2021 - Proceedings. - 2076-1465 .- 2219-5491. - 9789082797060 ; , s. 870-874
-
Conference paper (peer-reviewed)abstract
- Recent advances in robust self-calibration have made it possible to estimate microphone positions and at least partial sound source positions using ambient sound. However, there are limits on how well sound source paths can be recovered using state-of-the-art techniques. In this paper we develop and evaluate several techniques to extend partial and incomplete solutions. We present minimal solvers for sound source positioning using non-overlapping pairs of microphone positions and their respective time-difference measurements, and show how these new solvers can be used in a hypothesis and test setting. We also investigate techniques that exploit temporal smoothness of the sound source paths. The different techniques are evaluated on both real and synthetic data, and compared to several state-of-the-art techniques for time-difference-of-arrival multilateration.
|
|
| 6. |
- Ferranti, Luca, et al.
(author)
-
Homotopy Continuation for Sensor Networks Self-Calibration
- 2021
-
In: 29th European Signal Processing Conference, EUSIPCO 2021 - Proceedings. - 2219-5491. - 9789082797060 ; 2021-August, s. 1725-1729
-
Conference paper (peer-reviewed)abstract
- Given a sensor network, TDOA self-calibration aims at simultaneously estimating the positions of receivers and transmitters, and transmitters time offsets. This can be formulated as a system of polynomial equations. Due to the elevated number of unknowns and the nonlinearity of the problem, obtaining an accurate solution efficiently is nontrivial. Previous work has shown that iterative algorithms are sensitive to initialization and little noise can lead to failure in convergence. Hence, research has focused on algebraic techniques. Stable and efficient algebraic solvers have been proposed for some network configurations, but they do not work for smaller networks. In this paper, we use homotopy continuation to solve four previously unsolved configurations in 2D TDOA self-calibration, including a minimal one. As a theoretical contribution, we investigate the number of solutions of the new minimal configuration, showing this is much lower than previous estimates. As a more practical contribution, we also present new subminimal solvers, which can be used to achieve unique accurate solutions in previously unsolvable configurations. We demonstrate our solvers are stable both with clean and noisy data, even without nonlinear refinement afterwards. Moreover, we demonstrate the suitability of homotopy continuation for sensor network calibration problems, opening prospects to new applications.
|
|
| 7. |
- Ferranti, Luca, et al.
(author)
-
Multiple Offsets Multilateration: : A New Paradigm for Sensor Network Calibration with Unsynchronized Reference Nodes
- 2022
-
In: 2022 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2022 - Proceedings. - 1520-6149. - 9781665405409 ; 2022-May, s. 4958-4962
-
Conference paper (peer-reviewed)abstract
- Positioning using wave signal measurements is used in several applications, such as GPS systems, structure from sound and Wifi based positioning. Mathematically, such problems require the computation of the positions of receivers and/or transmitters as well as time offsets if the devices are unsynchronized. In this paper, we expand the previous state-of-the-art on positioning formulations by introducing Multiple Offsets Multilateration (MOM), a new mathematical framework to compute the receivers positions with pseudoranges from unsynchronized reference transmitters at known positions. This could be applied in several scenarios, for example structure from sound and positioning with LEO satellites. We mathematically describe MOM, determining how many receivers and transmitters are needed for the network to be solvable, a study on the number of possible distinct solutions is presented and stable solvers based on homotopy continuation are derived. The solvers are shown to be efficient and robust to noise both for synthetic and real audio data.
|
|
| 8. |
- Gummeson, Anna, et al.
(author)
-
Automatic Gleason grading of H&E stained microscopic prostate images using deep convolutional neural networks
- 2017
-
In: Medical Imaging 2017: Digital Pathology. - : SPIE. - 9781510607255 ; 10140
-
Conference paper (peer-reviewed)abstract
- Prostate cancer is the most diagnosed cancer in men. The diagnosis is confirmed by pathologists based on ocular inspection of prostate biopsies in order to classify them according to Gleason score. The main goal of this paper is to automate the classification using convolutional neural networks (CNNs). The introduction of CNNs has broadened the field of pattern recognition. It replaces the classical way of designing and extracting hand-made features used for classification with the substantially different strategy of letting the computer itself decide which features are of importance. For automated prostate cancer classification into the classes: Benign, Gleason grade 3, 4 and 5 we propose a CNN with small convolutional filters that has been trained from scratch using stochastic gradient descent with momentum. The input consists of microscopic images of haematoxylin and eosin stained tissue, the output is a coarse segmentation into regions of the four different classes. The dataset used consists of 213 images, each considered to be of one class only. Using four-fold cross-validation we obtained an error rate of 7.3%, which is significantly better than previous state of the art using the same dataset. Although the dataset was rather small, good results were obtained. From this we conclude that CNN is a promising method for this problem. Future work includes obtaining a larger dataset, which potentially could diminish the error margin.
|
|
| 9. |
- Örnhag, Marcus Valtonen, et al.
(author)
-
Minimal Solvers for Indoor UAV Positioning
- 2021
-
In: 2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR). - : IEEE COMPUTER SOC. - 1051-4651. - 9781728188089 ; , s. 1136-1143
-
Conference paper (peer-reviewed)abstract
- In this paper we consider a collection of relative pose problems which arise naturally in applications for visual indoor navigation using unmanned aerial vehicles (UAVs). We focus on cases where additional information from an onboard IMU is available and thus provides a partial extrinsic calibration through the gravitational vector. The solvers are designed for a partially calibrated camera, for a variety of realistic indoor scenarios, which makes it possible to navigate using images of the ground floor. Current state-of-the-art solvers use more general assumptions, such as using arbitrary planar structures; however, these solvers do not yield adequate reconstructions for real scenes, nor do they perform fast enough to be incorporated in real-time systems. We show that the proposed solvers enjoy better numerical stability, are faster, and require fewer point correspondences, compared to state-of-the-art approaches. These properties are vital components for robust navigation in real-time systems, and we demonstrate on both synthetic and real data that our method outperforms other solvers, and yields superior motion estimation(1).
|
|
