A5100381073- Anomaly Detection Techniques and Applications
- Fault Detection and Control Systems
- Time Series Analysis and Forecasting
- Advanced Statistical Methods and Models
- Advanced Neural Network Applications
- AI in cancer detection
- Network Security and Intrusion Detection
- Water Systems and Optimization
- Advanced Vision and Imaging
- Radiomics and Machine Learning in Medical Imaging
- Digital Imaging for Blood Diseases
- 3D Shape Modeling and Analysis
- Adversarial Robustness in Machine Learning
- Privacy-Preserving Technologies in Data
- Computer Graphics and Visualization Techniques
- Medical Image Segmentation Techniques
- CCD and CMOS Imaging Sensors
- Anatomy and Medical Technology
- EEG and Brain-Computer Interfaces
- Advanced Statistical Process Monitoring
- Automated Road and Building Extraction
- Advanced Clustering Algorithms Research
- Particle Detector Development and Performance
- Data-Driven Disease Surveillance
- Machine Learning and Data Classification
University of Turku
2024
Northwestern Polytechnical University
2019-2024
Guangzhou University
2021-2024
Shanghai University
2024
Kunming University of Science and Technology
2024
Tianjin Normal University
2024
China National Center for Food Safety Risk Assessment
2024
Peking University
2024
Harbin Institute of Technology
2023-2024
Ningbo Institute of Industrial Technology
2024
Traditional outlier detection methods create a model for data and then label as outliers objects that deviate significantly from this model. However, when dat has many outliers, also pollute the The becomes unreliable, thus rendering most detectors to become ineffective. To solve problem, we propose mean-shift detector. This detector employs technique modify cancel bias caused by outliers. replaces every object mean of its k-nearest neighbors which essentially removes effect before...
Novel view synthesis with sparse inputs is a challenging problem for neural radiance fields (NeRF). Recent efforts alleviate this challenge by introducing external supervision, such as pre-trained models and extra depth signals, or using non-trivial patch-based rendering. In paper, we present Frequency regularized NeRF (FreeNeRF), surprisingly simple baseline that outperforms previous methods minimal modifications to plain NeRF. We analyze the key challenges in few-shot rendering find...
Outlier detection is a fundamental issue in data mining and machine learning. Most methods calculate outlier score for each object then threshold the scores to detect outliers. widely used thresholding techniques are based on statistics like standard deviation around mean, median absolute interquartile range. Unfortunately, these can be significantly biased because of presence outliers when calculating statistics. This makes their use inaccurate. To overcome this problem, we propose...
A previous study has examined the overall cancer statistics. However, more detailed statistics regarding liver have not been provided. We evaluated incidence and mortality trends of intrahepatic bile duct in United States from 1975 to 2017 based on data Surveillance, Epidemiology, End Results (SEER) database.Age, gender, race, metastasis, tumor site, grade patients were extracted SEER database. Codes C22.0 C22.1 International Classification Disease for Oncology applied identify with...
Classification of interbeat interval time-series which fluctuates in an irregular and complex manner is very challenging. Typically, entropy methods are employed to quantify the complexity for classifying. Traditional focus on frequency distribution all observations a time-series. This requires relatively long with at least couple thousands data points, limits their usages practical applications. The also sensitive parameter settings. In this paper, we propose conceptually new approach...
Over the decades, traditional outlier detectors have ignored group-level factor when calculating scores for objects in data by evaluating only object-level factor, failing to capture collective outliers. To mitigate this issue, we present a framework called neighborhood representative (NR), which empowers all existing efficiently detect outliers, including while maintaining their computational integrity. It achieves selecting objects, scoring these then applies score of its objects. Without...
We present EmerNeRF, a simple yet powerful approach for learning spatial-temporal representations of dynamic driving scenes. Grounded in neural fields, EmerNeRF simultaneously captures scene geometry, appearance, motion, and semantics via self-bootstrapping. hinges upon two core components: First, it stratifies scenes into static fields. This decomposition emerges purely from self-supervision, enabling our model to learn general, in-the-wild data sources. Second, parameterizes an induced...
Panoramic X-ray (PX) provides a 2D picture of the patient's mouth in panoramic view to help dentists observe invisible disease inside gum. However, it limited information compared with cone-beam computed tomography (CBCT), another dental imaging method that generates 3D oral cavity but more radiation dose and higher price. Consequently, is great interest reconstruct structure from image, which can greatly explore application surgeries. In this paper, we propose framework, named Oral-3D,...
Outlier ensemble is an important methodology for improving outlier detection, but faces severe challenges in unsupervised settings. Unlike traditional ensembles which revised scores by considering only the values of from multiple detectors, we present a novel Regional Ensemble (RE). RE combines objects and detectors simultaneously takes into consideration both distribution these scores. specifically enhances score given object using neighboring object, under assumption that majority are...
Deep learning-based sequence models are extensively employed in Time Series Anomaly Detection (TSAD) tasks due to their effective sequential modeling capabilities.However, the ability of TSAD is limited by two key challenges: (i) model long-range dependency and (ii) generalization issue presence non-stationary data.To tackle these challenges, an anomaly detector that leverages selective state space known for its proficiency capturing long-term dependencies across various domains...
The enhanced x-ray timing and polarimetry mission (eXTP) is a flagship observatory for timing, spectroscopy developed by an international consortium. Thanks to its very large collecting area, good spectral resolution unprecedented capabilities, eXTP will explore the properties of matter propagation light in most extreme conditions found universe. will, addition, be powerful observatory. continuously monitor sky, enable multi-wavelength multi-messenger studies. currently phase B, which...
Extracting road segments by averaging GPS trajectories is very challenging. Most existing strategies suffer from high complexity, poor accuracy, or both. For example, finding the optimal mean for a set of sequences known to be NP-hard, whereas using Medoid compromises quality. In this paper, we introduce three extremely fast and practical methods extract segment trajectories. The first analyze descriptors then use either simple linear model more complex curvy depending on an angle criterion....
Few-shot learning is an established topic in natural images for years, but few work attended to histology images, which of high clinical value since well-labeled datasets and rare abnormal samples are expensive collect. Here, we facilitate the study few-shot by setting up three cross-domain tasks that simulate real clinics problems. To enable label-efficient better generalizability, propose incorporate contrastive (CL) with latent augmentation (LA) build a system. CL learns useful...
Accurate brain tumor segmentation from Magnetic Resonance Imaging (MRI) is desirable to joint learning of multimodal images. However, in clinical practice, it not always possible acquire a complete set MRIs, and the problem missing modalities causes severe performance degradation existing methods. In this work, we present first attempt exploit Transformer for that robust any combinatorial subset available modalities. Concretely, propose novel Medical (mmFormer) incomplete with three main...
Most conventional time series outlier detection techniques exhibit limitations for variable-length series. To tackle this problem, we propose HK-index to extract features from any detectors. identifies the longest subseries which satisfies conditions by a pre-defined parameter <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$k$</tex> . The length of found is then used as feature. novelty core idea use information present whole series, ignoring...