Timo Ojala

Fabi Prezja, Leevi Annala, Sampsa Kiiskinen et al.

In routine colorectal cancer management, histologic samples stained with hematoxylin and eosin are commonly used. Nonetheless, their potential for defining objective biomarkers for patient stratification and treatment selection is still being explored. The current gold standard relies on expensive and time-consuming genetic tests. However, recent research highlights the potential of convolutional neural networks (CNNs) in facilitating the extraction of clinically relevant biomarkers from these readily available images. These CNN-based biomarkers can predict patient outcomes comparably to golden standards, with the added advantages of speed, automation, and minimal cost. The predictive potential of CNN-based biomarkers fundamentally relies on the ability of convolutional neural networks (CNNs) to classify diverse tissue types from whole slide microscope images accurately. Consequently, enhancing the accuracy of tissue class decomposition is critical to amplifying the prognostic potential of imaging-based biomarkers. This study introduces a hybrid Deep and ensemble machine learning model that surpassed all preceding solutions for this classification task. Our model achieved 96.74% accuracy on the external test set and 99.89% on the internal test set. Recognizing the potential of these models in advancing the task, we have made them publicly available for further research and development.

Fabi Prezja, Leevi Annala, Sampsa Kiiskinen et al.

Diagnosing knee joint osteoarthritis (KOA), a major cause of disability worldwide, is challenging due to subtle radiographic indicators and the varied progression of the disease. Using deep learning for KOA diagnosis requires broad, comprehensive datasets. However, obtaining these datasets poses significant challenges due to patient privacy concerns and data collection restrictions. Additive data augmentation, which enhances data variability, emerges as a promising solution. Yet, it's unclear which augmentation techniques are most effective for KOA. This study explored various data augmentation methods, including adversarial augmentations, and their impact on KOA classification model performance. While some techniques improved performance, others commonly used underperformed. We identified potential confounding regions within the images using adversarial augmentation. This was evidenced by our models' ability to classify KL0 and KL4 grades accurately, with the knee joint omitted. This observation suggested a model bias, which might leverage unrelated features for classification currently present in radiographs. Interestingly, removing the knee joint also led to an unexpected improvement in KL1 classification accuracy. To better visualize these paradoxical effects, we employed Grad-CAM, highlighting the associated regions. Our study underscores the need for careful technique selection for improved model performance and identifying and managing potential confounding regions in radiographic KOA deep learning.

Fabi Prezja, Leevi Annala, Sampsa Kiiskinen et al.

Knee Osteoarthritis (KOA), a leading cause of disability worldwide, is challenging to detect early due to subtle radiographic indicators. Diverse, extensive datasets are needed but are challenging to compile because of privacy, data collection limitations, and the progressive nature of KOA. However, a model capable of projecting genuine radiographs into different OA stages could augment data pools, enhance algorithm training, and offer pre-emptive prognostic insights. In this study, we trained a CycleGAN model to synthesize past and future stages of KOA on any genuine radiograph. The model was validated using a Convolutional Neural Network that was deceived into misclassifying disease stages in transformed images, demonstrating the CycleGAN's ability to effectively transform disease characteristics forward or backward in time. The model was particularly effective in synthesizing future disease states and showed an exceptional ability to retroactively transition late-stage radiographs to earlier stages by eliminating osteophytes and expanding knee joint space, signature characteristics of None or Doubtful KOA. The model's results signify a promising potential for enhancing diagnostic models, data augmentation, and educational and prognostic usage in healthcare. Nevertheless, further refinement, validation, and a broader evaluation process encompassing both CNN-based assessments and expert medical feedback are emphasized for future research and development.

Fabi Prezja, Leevi Annala, Sampsa Kiiskinen et al.

Knee-Joint Osteoarthritis (KOA) is a prevalent cause of global disability and is inherently complex to diagnose due to its subtle radiographic markers and individualized progression. One promising classification avenue involves applying deep learning methods; however, these techniques demand extensive, diversified datasets, which pose substantial challenges due to medical data collection restrictions. Existing practices typically resort to smaller datasets and transfer learning. However, this approach often inherits unnecessary pre-learned features that can clutter the classifier's vector space, potentially hampering performance. This study proposes a novel paradigm for improving post-training specialized classifiers by introducing adaptive variance thresholding (AVT) followed by Neural Architecture Search (NAS). This approach led to two key outcomes: an increase in the initial accuracy of the pre-trained KOA models and a 60-fold reduction in the NAS input vector space, thus facilitating faster inference speed and a more efficient hyperparameter search. We also applied this approach to an external model trained for KOA classification. Despite its initial performance, the application of our methodology improved its average accuracy, making it one of the top three KOA classification models.

Jacqueline Yau, Katherine J. Mimnaugh, Evan G. Center et al.

Cybersickness poses a serious challenge for users of virtual reality (VR) technology. Consequently, there has been significant effort to track its occurrence during VR use with passive measures like brain activity recorded through electroencephalogram (EEG). To classify cybersickness accurately, including in real time, machine learning algorithms which can extract meaningful signals from the rest of the brain data will be required. However, EEG datasets are typically very small and very high in variability between participants, which makes building effective models extremely challenging. To address these concerns, we first introduce a framework for neural networks which has subject-adaptive training with calibration and interpretation for classification given limited and imbalanced EEG data. Which features the models determine are most useful can be visualized by plotting interpretability maps from integrated gradients and class activation. The framework is demonstrated here with convolutional neural networks and transformer models. Using a set of brain data recorded with EEG while participants viewed a stimulus in VR designed to elicit cybersickness, we show which spatio-temporal EEG features (from electrodes and time steps) were most important for discomfort classification. Across 12 runs of our framework with three different neural networks over multiple random seeds, the models consistently pointed to the same scalp locations as having patterns of brain data that were the most helpful in determining whether or not a sample of EEG data belonged to someone who was experiencing cybersickness. These results help clarify a hidden pattern in other related research and can be used as tagged features for better real-time cybersickness classification with EEG in the future. We provide our code at [anonymized] to enable feature interpretation across different neural network architectures.

Markku Suomalainen, Basak Sakcak, Adhi Widagdo et al.

We propose unwinding the rotations experienced by the user of an immersive telepresence robot to improve comfort and reduce VR sickness of the user. By immersive telepresence we refer to a situation where a 360\textdegree~camera on top of a mobile robot is streaming video and audio into a head-mounted display worn by a remote user possibly far away. Thus, it enables the user to be present at the robot's location, look around by turning the head and communicate with people near the robot. By unwinding the rotations of the camera frame, the user's viewpoint is not changed when the robot rotates. The user can change her viewpoint only by physically rotating in her local setting; as visual rotation without the corresponding vestibular stimulation is a major source of VR sickness, physical rotation by the user is expected to reduce VR sickness. We implemented unwinding the rotations for a simulated robot traversing a virtual environment and ran a user study (N=34) comparing unwinding rotations to user's viewpoint turning when the robot turns. Our results show that the users found unwound rotations more preferable and comfortable and that it reduced their level of VR sickness. We also present further results about the users' path integration capabilities, viewing directions, and subjective observations of the robot's speed and distances to simulated people and objects.

Matti Pouke, Katherine J. Mimnaugh, Alexis Chambers et al.

This paper identifies and confirms a perceptual phenomenon: when users interact with simulated objects in a virtual environment where the users' scale deviates greatly from normal, there is a mismatch between the object physics they consider realistic and the object physics that would be correct at that scale. We report the findings of two studies investigating the relationship between perceived realism and a physically accurate approximation of reality in a virtual reality experience in which the user has been scaled by a factor of ten. Study 1 investigated perception of physics when scaled-down by a factor of ten, whereas Study 2 focused on enlargement by a similar amount. Studies were carried out as within-subjects experiments in which a total of 84 subjects performed simple interaction tasks with objects under two different physics simulation conditions. In the true physics condition, the objects, when dropped and thrown, behaved accurately according to the physics that would be correct at that either reduced or enlarged scale in the real world. In the movie physics condition, the objects behaved in a similar manner as they would if no scaling of the user had occurred. We found that a significant majority of the users considered the movie physics condition to be the more realistic one. However, at enlarged scale, many users considered true physics to match their expectations even if they ultimately believed movie physics to be the realistic condition. We argue that our findings have implications for many virtual reality and telepresence applications involving operation with simulated or physical objects in abnormal and especially small scales.

Matti Pouke, Katherine J. Mimnaugh, Timo Ojala et al.

This paper identifies a new phenomenon: when users interact with simulated objects in a virtual environment where the user is much smaller than usual, there is a mismatch between the object physics that they expect and the object physics that would be correct at that scale. We report the findings of our study investigating the relationship between perceived realism and a physically accurate approximation of reality in a virtual reality experience in which the user has been scaled down by a factor of ten. We conducted a within-subjects experiment in which 44 subjects performed a simple interaction task with objects under two different physics simulation conditions. In one condition, the objects, when dropped and thrown, behaved accurately according to the physics that would be correct at that reduced scale in the real world, our true physics condition. In the other condition, the movie physics condition, the objects behaved in a similar manner as they would if no scaling of the user had occurred. We found that a significant majority of the users considered the latter condition to be the more realistic one. We argue that our findings have implications for many virtual reality and telepresence applications involving operation with simulated or physical objects in small scales.