Tag: focal loss

Researchers developed SegAnyNeuron, a feature modeling framework that improves three-dimensional neuronal image segmentation across diverse microscopy modalities by encoding local ...

Researchers introduced CSwinUKAN, an enhanced U-Net model integrating Swin Transformers and Kolmogorov–Arnold Networks to improve crack segmentation recall in operating ...

Researchers developed a confidence-gated convolutional neural network and bidirectional recurrent model that enhances rare arrhythmia classification on electrocardiogram signals while ...

A new dual-branch network combines raw ECG signals with wavelet-transformed frequency features and adaptive attention, reaching 99.38% accuracy on the ...

A hybrid ResNet50–VGG16 model with Coordinate Attention and focal loss reached a macro F1-score of 0.718 on the 22-class GastroVision ...

A deep-learning framework classified sleep stages from single-channel EEG with reported accuracies of 84.5% to 87.0% across three public datasets, ...

A new deep learning architecture called FrDPalmNet combines fractional calculus, dual attention, and densely connected blocks to push palmprint-based biometric ...

Researchers in Qatar have built a two-stage AI pipeline that predicts clinical keywords from chest X-rays and then generates full ...

A new multi-scale deep learning framework called Rain-PostEFNet substantially improves numerical weather prediction rainfall forecasts, with the largest gains in ...

Researchers have introduced Light-wDEF, a resource-efficient and fully reproducible framework that enabled a fair comparison of ten visual deepfake detectors ...