Export Citations
Save this search
Please login to be able to save your searches and receive alerts for new content matching your search criteria.
- research-articleMarch 2024
ConvGeN: A convex space learning approach for deep-generative oversampling and imbalanced classification of small tabular datasets
AbstractOversampling is commonly used to improve classifier performance for small tabular imbalanced datasets. State-of-the-art linear interpolation approaches can be used to generate synthetic samples from the convex space of the minority class. ...
Highlights- Deep adversarial learning isn’t apt for oversampling on small tabular imbalanced data.
- ConvGeN learns appropriate convex coefficients from each minority data neighborhood.
- ConvGeN is the first deep learning architecture for ...
- research-articleMarch 2024
H-CapsNet: A capsule network for hierarchical image classification
AbstractIn this paper, we present H-CapsNet, a capsule network for hierarchical image classification. Our network makes use of the natural capacity of CapsNets (capsule networks) to capture hierarchical relationships. Thus, our network is such that each ...
Highlights- Propose H-CapsNet for hierarchical classification with dedicated capsules per level.
- Enforce hierarchical consistency with label-tree guided modified hinge-loss.
- Adjust training parameters dynamically to balance hierarchical level ...
- research-articleMarch 2024
Joint Feature Generation and Open-set Prototype Learning for generalized zero-shot open-set classification
AbstractIn generalized zero-shot classification, test samples can belong to either seen or unseen classes. However, in real-world situations, there may be many open-set samples in the test set where neither visual nor semantic representations of the ...
Highlights- A new problem of the generalized zero-shot open-set classification is proposed.
- Feature generation model and open-set prototype learning are unified for GZSOSC.
- A novel open-set prototype learning method is proposed.
- The ...
- research-articleMarch 2024
When IC meets text: Towards a rich annotated integrated circuit text dataset
AbstractAutomated Optical Inspection (AOI) is a process that uses cameras to autonomously scan printed circuit boards for quality control. Text is often printed on chip components, and it is crucial that this text is correctly recognized during AOI, as ...
Highlights- We present a large-scale ICText dataset labeled with multi-label quality attributes.
- We propose a two-stage AGCL loss to reweight the loss in a plug-and-play fashion.
- AGCL trains from easy to hard samples given low contrast, blurry,...
- research-articleMarch 2024
Scalable and accurate subsequence transform for time series classification
AbstractTime series classification using phase-independent subsequences called shapelets is one of the best approaches in the state of the art. This approach is especially characterized by its interpretable property and its fast prediction time. However, ...
Highlights- We introduce the core shapelet recognition task.
- We claim that time series classification by shapelets is a core shapelet recognition task.
- We propose the SAST method to successfully performs the core shapelet recognition task in O ...
-
- research-articleMarch 2024
PWDformer: Deformable transformer for long-term series forecasting
AbstractLong-term forecasting is of paramount importance in numerous scenarios, including predicting future energy, water, and food consumption. For instance, extreme weather events and natural disasters can profoundly impact infrastructure operations ...
Highlights
- Empirical Evaluation of High-Performance Time Series Forecasting Techniques for Single and Multivariate Data.
- Our deformable mechanism-based module significantly enhances the performance of the model.
- Our position-sensitive ...
- research-articleMarch 2024
Learning conditional variational autoencoders with missing covariates
AbstractConditional variational autoencoders (CVAEs) are versatile deep latent variable models that extend the standard VAE framework by conditioning the generative model with auxiliary covariates. The original CVAE model assumes that the data samples ...
Graphical abstractDisplay Omitted
Highlights- An improved learning method for conditional VAEs and Gaussian process prior VAEs.
- The method is designed for non-temporal, temporal, and longitudinal data.
- Used an amortised variational distribution for learning missing auxiliary ...
- research-articleMarch 2024
DBN-Mix: Training dual branch network using bilateral mixup augmentation for long-tailed visual recognition
AbstractThere is growing interest in the challenging visual perception task of learning from long-tailed class distributions. The extreme class imbalance in the training dataset biases the model to prefer recognizing majority class data over minority ...
Highlights- The proposed bilateral mixup improves representation learning for minority class.
- Class-wise temperature scaling effectively mitigates the bias in classifier.
- Dual-branch network with two main ideas achieves state-of-the-art ...
- research-articleMarch 2024
F-SCP: An automatic prompt generation method for specific classes based on visual language pre-training models
AbstractThe zero-shot classification performance of large-scale vision-language pre-training models (e.g., CLIP, BLIP and ALIGN) can be enhanced by incorporating a prompt (e.g., “a photo of a [CLASS]”) before the class words. Modifying the prompt ...
Highlights- The prompt greatly affects visual language pre-training models’ classification performance.
- Classes with lower accuracy than others and similar classes are specific classes.
- We filter out the poorly performing classes for separate ...
- research-articleMarch 2024
Learning a target-dependent classifier for cross-domain semantic segmentation: Fine-tuning versus meta-learning
Highlights- The introduction of a target-dependent classifier for cross-domain semantic segmentation able to better fit with the target domain features even under imperfect domain alignment.
- An innovative linkage of meta-learning with domain ...
Recently proposed domain adaptation arts have dominated the field of cross-domain semantic segmentation by operating domain manifolds alignment and learning an optimal joint hypothesis (joint-domain classifier) for both source and target domains. ...
- research-articleMarch 2024
View-coherent correlation consistency for semi-supervised semantic segmentation
AbstractSemi-supervised semantic segmentation needs rich and robust supervision for unlabeled data. However, promoting or punishing feature similarities with vanilla contrastive learning can be unreliable for semi-supervised semantic segmentation: pixel ...
Highlights- This work investigates semi-supervised semantic segmentation.
- Correlation consistency to combine consistency and contrastive learning.
- New pipeline with view-coherent data augmentation.
- State-of-the-art results on multiple ...
- research-articleMarch 2024
Learning implicit labeling-importance and label correlation for multi-label feature selection with streaming labels
AbstractMulti-label feature selection plays an increasingly important role in alleviating the high dimensionality of multi-label learning tasks. Most extant methods posit that the learning task is performed in an environment where the label space is ...
Highlights- By jointly learning the implicit labeling-importance for streaming labels and the impact of label correlation on feature importance, a novel framework is established to figure out multi-label feature selection with dynamic streaming ...
- research-articleMarch 2024
A linear transportation L p distance for pattern recognition
- Oliver M. Crook,
- Mihai Cucuringu,
- Tim Hurst,
- Carola-Bibiane Schönlieb,
- Matthew Thorpe,
- Konstantinos C. Zygalakis
AbstractThe transportation L p distance, denoted TL p, has been proposed as a generalisation of Wasserstein W p distances motivated by the property that it can be applied directly to colour or multi-channelled images, as well as multivariate time-series ...
Highlights- We linearise a functional type optimal transport distance.
- We demonstrate that this linearisation is computationally efficient and allows for off-the-shelf data analysis tools.
- We propose extensions to higher orders.
- In ...
- research-articleMarch 2024
Time pattern reconstruction for classification of irregularly sampled time series
AbstractIrregularly Sampled Time Series (ISTS) include partially observed feature vectors caused by the lack of temporal alignment across dimensions and the presence of variable time intervals. Especially in medical applications, because patients’ ...
Highlights- Classifying irregularly sampled time series is crucial in real-world applications.
- Incorporate time-depended mapping into time series classification.
- Extract time patterns of irregularly sampled time series.
- Propose a plug-in ...
- research-articleMarch 2024
A general elevating framework for label noise filters
AbstractIn real applications, label noise has a great influence on data modeling. As one kind of label noise treatment method, noise filter has attracted extensive attention recently. The existing filters perform well in dealing with label noise ...
Highlights- The proposed noise filtering framework can alleviate the phenomenon of over-cleaning.
- The performance of most existing noise filters can be improved with this framework.
- The metric recommended in this paper can guide the research ...
- research-articleMarch 2024
DCapsNet: Deep capsule network for human activity and gait recognition with smartphone sensors
Highlights- A deep capsule network model named DCapsNet is proposed to automatically extract the activity or gait features from sensors and classify them.
- The proposed model combines a set of convolutional layers and capsule network.
- The set ...
Recently, deep neural networks are used to recognize human activity/gait through mobile sensors which have attracted a great attention. Although the existing deep neural networks that perform automatic feature extraction have achieved desirable ...
- research-articleMarch 2024
Fairness in face presentation attack detection
AbstractFace recognition (FR) algorithms have been proven to exhibit discriminatory behaviors against certain demographic and non-demographic groups, raising ethical and legal concerns regarding their deployment in real-world scenarios. Despite the ...
Highlights- A novel Combined Attribute Annotated PAD Dataset including seven attribute labels.
- A metric to jointly represent the absolute PAD performance and the PAD fairness.
- Detailed fairness assessments reveals unfairness induced by data ...
- research-articleFebruary 2024
Pseudo-label estimation via unsupervised Identity Link Prediction for one-shot person Re-Identification
AbstractIn this paper, we propose an unsupervised identity link prediction (ILP) method for label estimation in one-shot person Re-ID. ILP aims to relax the constraints of labeled samples and group a set of unlabeled pedestrians by their potential ...
Highlights- We propose an unsupervised identity link prediction (ILP) method for label estimation in one-shot person Re-ID.
- We construct an Identity Link Subgraph (ILS) by a two-step procedure to describe the link relationship between nodes and ...
- research-articleFebruary 2024
Introducing instance label correlation in multiple instance learning. Application to cancer detection on histopathological images
AbstractIn the last years, the weakly supervised paradigm of multiple instance learning (MIL) has become very popular in many different areas. A paradigmatic example is computational pathology, where the lack of patch-level labels for whole-slide images ...
Highlights- We introduce instance label correlation for GP-based MIL models.
- To do so, we develop a novel coupling term inspired by the stat-physics Ising model.
- When it vanishes, we recover the equations of a well-known GP-based MIL method.
- research-articleFebruary 2024
Phase Randomization: A data augmentation for domain adaptation in human action recognition
AbstractHuman action recognition models often suffer from achieving both accurate recognition and subject independence when the amount of training data is limited. In this paper, we propose a data-efficient domain adaptation approach to learning a ...
Highlights- Data-efficient domain adaptation for subject-agnostic action recognition.
- Phase Randomization data augmentation to disentangle action and subject features.
- Extensive experiments on two action recognition benchmark tasks.