Abstract: Unsupervised domain adaptation (UDA) techniques are vital for semantic segmentation in geosciences, effectively utilizing remote sensing imagery across diverse domains. However, most ...

We propose MaskCut approach to generate pseudo-masks for multiple objects in an image. CutLER can learn unsupervised object detectors and instance segmentors solely on ImageNet-1K. CutLER exhibits ...

Unsupervised multilabel image segmentation (colour, grayscale, multichannel) via the Potts model — also known as the piecewise-constant Mumford-Shah model or the ℓ⁰ gradient model. Solvers for 1-D ...

Neural networks are powerful tools for processing visual inputs, but precisely how this processing is performed remains unclear. We introduce a recurrent neural network that can perform simple image ...

Unsupervised domain adaptation (UDA) aims to adapt a model learned from the source domain to the target domain. Thus, the model can obtain transferable knowledge even in target domain that does not ...

Laser speckle contrast imaging (LSCI) is a full-field, high spatiotemporal resolution and low-cost optical technique for measuring blood flow, which has been successfully used for neurovascular ...

1 Shanghai Key Laboratory of Multidimensional Information Processing, East China Normal University, Shanghai, China. 2 Department of Medical Ultrasound, Fudan University Shanghai Cancer Center; ...

Abstract: The usage of convolutional neural networks (CNNs) for unsupervised image segmentation was investigated in this study. Similar to supervised image segmentation, the proposed CNN assigns ...