Sorry, you need to enable JavaScript to visit this website.

Ischemic stroke lesions (ISL) segmentation aids clinicians in the diagnosis of stroke in acute care units. But, a generalized segmentation model requires data from various patients. However considering the data privacy, the patient's data is not available for centralized training. The Federated Learning (FL) framework overcomes this, but in FL, semantic segmentation is challenging due to the complex model, adversarial training, and non-independent and identically distributed dataset.

Recent advancements in learning algorithms have demonstrated that the sharpness of the loss surface is an effective measure for improving the generalization gap. Building upon this concept, Sharpness-Aware Minimization (SAM) was proposed to enhance model generalization and achieved state-of-the-art performance. SAM consists of two main steps, the weight perturbation step and the weight updating step. However, the perturbation in SAM is determined by only the gradient of the training loss, or cross-entropy loss.

In this work we utilize different open-source datasets and deep learning models for detecting objects from image data captured by a mobile mapping system integrating the multi-camera Ladybug 5+ in an urban area. In our experiments
we exploit sets of pre-trained models and models trained via transfer learning techniques with available open source datasets for object detection, semantic-, instance-, and panoptic segmentation. Tests with the trained models are performed with image data from the Ladybug 5+ camera.

Point Clouds (PCs) have gained significant attention due to their usage in diverse application domains, notably virtual and augmented reality. While PCs excel in providing detailed 3D visualization, this typically requires millions of points which must be efficiently coded for real-world deployment, notably storage and streaming. Recently, learning-based coding solutions have been adopted, notably in the JPEG Pleno Point Coding (PCC) standard, which uses a coding model with millions of model parameters.

Counting repetitive actions in long untrimmed videos is a challenging task that has many applications such as rehabilitation. State-of-the-art methods predict action counts by first generating a temporal self-similarity matrix (TSM) from the sampled frames and then feeding the matrix to a predictor network. The self-similarity matrix, however, is not an optimal input to a network since it discards too much information from the frame-wise embeddings.

Low-rank Deconvolution (LRD) has appeared as a new multi-dimensional representation model that enjoys important efficiency and flexibility properties. In this work we ask ourselves if this analytical model can compete against Deep Learning (DL) frameworks like Deep Image Prior (DIP) or Blind-Spot Networks (BSN) and other classical methods in the task of signal restoration. More specifically, we propose to extend LRD with differential regularization.

Low-rank Deconvolution (LRD) has appeared as a new multi-dimensional representation model that enjoys important efficiency and flexibility properties. In this work we ask ourselves if this analytical model can compete against Deep Learning (DL) frameworks like Deep Image Prior (DIP) or Blind-Spot Networks (BSN) and other classical methods in the task of signal restoration. More specifically, we propose to extend LRD with differential regularization.

3D scene understanding is crucial for facilitating seamless interaction between digital devices and the physical world. Real-time capturing and processing of the 3D scene are essential for achieving this seamless integration. While existing approaches typically separate acquisition and processing for each frame, the advent of resolution-scalable 3D sensors offers an opportunity to overcome this paradigm and fully leverage the otherwise wasted acquisition time to initiate processing.

This study introduces an innovative framework for accurate colon segmentation in abdomen CT scans, addressing the unique challenges of this task. Our architecture enhances well-established 2D segmentation models by incorporating 3D contextual information through a novel method that generates an attention map for a given slice by considering its neighboring slices. This approach achieves effective colon segmentation without complex 3D convolutional neural networks (CNNs) or Long Short-Term Memory (LSTM) networks by combining 2D CNNs.

Deep learning in image classification has achieved remarkable success but at the cost of high resource demands. Model compression through automatic joint pruning-quantization addresses this issue, yet most existing techniques overlook a critical aspect: layer correlations. These correlations are essential as they expose redundant computations across layers, and leveraging them facilitates efficient design space exploration. This study employs Graph Neural Networks (GNN) to learn these inter-layer relationships, thereby optimizing the pruning-quantization strategy for the targeted model.