Masked Autoencoder
PulseAugur coverage of Masked Autoencoder — every cluster mentioning Masked Autoencoder across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
FastMap framework enables real-time semantic map completion for robots
Researchers have developed FastMap, a novel two-stage framework designed for real-time semantic map completion in indoor robot navigation. This system utilizes a BitVAE to compress semantic map patches into compact bitw…
-
New SARA attack bypasses Vision Transformer privacy defenses
A new research paper details a feature inversion attack called SARA that can reconstruct input images from Vision Transformer (ViT) embeddings transmitted in split-inference systems. The attack demonstrates that token s…
-
New framework adapts AI models for material recognition from sparse visual data
A new framework called Sparse Surface Understanding Framework (SSUF) has been developed to improve material recognition from incomplete visual data. SSUF adapts four pre-trained architectures—ConvAE, ViT, Swin Transform…
-
CDG-MAE uses diffusion models for synthetic views in computer vision
Researchers have developed CDG-MAE, a novel self-supervised learning method for computer vision that utilizes synthetic views generated by diffusion models. This approach addresses the challenge of acquiring diverse tra…
-
Foundation model pretraining strategies impact retinal imaging transferability
A new arXiv paper explores how different pretraining strategies for foundation models impact their effectiveness when transferred to ultra-widefield retinal imaging tasks. Researchers compared Vision Transformer encoder…
-
New SPFM-Net framework targets invisible watermarks with Mamba architecture
Researchers have developed SPFM-Net, a novel framework designed to attack invisible watermarks in images. This system utilizes a semantic-prior-guided and frequency-constrained Mamba architecture to effectively remove w…
-
Visual MAE adapted for time series anomaly detection
Researchers have developed VAN-AD, a novel framework for time series anomaly detection that adapts a visual Masked Autoencoder (MAE) pretrained on ImageNet. This approach aims to improve generalization capabilities acro…
-
Masked Autoencoder learns steel defect recognition with 91.3% accuracy
Researchers have developed a novel unsupervised method for recognizing steel surface defects using a Transformer-based Masked Autoencoder. This approach learns representations from abundant unlabeled images by masking 7…
-
New ProsMAE framework enhances histopathology representation learning
Researchers have developed ProsMAE, a novel multi-source Masked Autoencoder framework designed for histopathology representation learning. This approach utilizes tiles from diverse datasets like PANDA, CAMELYON17, and B…
-
SleepMaMi: Novel Foundation Model Integrates Sleep Architecture and Biosignals
Researchers have developed SleepMaMi, a novel sleep foundation model designed to integrate both long-term sleep architecture and fine-grained biosignal analysis. This model employs a hierarchical dual-encoder structure,…
-
CUPID deepfake detector uses UV maps and MAE for interpretable analysis
Researchers have developed CUPID, a novel deepfake detection method that reconstructs UV texture maps from 3D face models and utilizes Masked Autoencoders (MAE) for analysis. This approach does not require deepfake vide…
-
New framework cuts medical image annotation effort using self-supervision
Researchers have developed a new framework called XSSR to reduce the effort needed for annotating medical images across different domains. The method uses a self-supervised approach with a Masked Autoencoder to learn fr…
-
New CA-LIG framework enhances Transformer model explainability
Researchers have developed a new framework called Context-Aware Layer-wise Integrated Gradients (CA-LIG) to improve the explainability of Transformer models. This framework offers a unified, hierarchical approach that c…
-
New framework boosts medical image classification with dual model approach
Researchers have developed a new deep learning framework for medical image classification that combines self-supervised and transfer learning techniques. The approach utilizes two ConvNeXt-Tiny models, one pre-trained o…