Eurosat
PulseAugur coverage of Eurosat — every cluster mentioning Eurosat across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
AI research shows language prompts vary in usefulness after visual adaptation
Researchers have investigated the effectiveness of language descriptions in source-free cross-domain few-shot learning (SF-CDFSL). Their study, focusing on datasets like EuroSAT and CropDisease, reveals two distinct reg…
-
EfficientViT-M2 leads in robust onboard satellite image classification
A comparative study evaluated 14 different computer vision models, including various Vision Transformer (ViT) architectures, for onboard satellite image classification in Earth observation tasks. The research focused on…
-
Spiking Neural Networks enhanced for remote sensing OOD detection
Researchers have developed a novel method to improve out-of-distribution (OOD) detection in Spiking Neural Networks (SNNs) for remote sensing applications. Their approach utilizes a spiking pseudo-ensemble, where multip…
-
QScheduler algorithm enables adaptive on-device AI training on microcontrollers
Researchers have developed QScheduler, an adaptive algorithm designed to optimize on-device training for microcontrollers equipped with Neural Processing Units (NPUs). This method estimates gradients using only forward …
-
New CCPL method enhances few-shot CLIP adaptation
Researchers have developed a new method called Concept-Constrained Prompt Learning (CCPL) to improve the adaptation of CLIP models for few-shot learning tasks. This framework uses regularization to anchor learnable clas…
-
In-sensor computing boosts satellite Earth observation efficiency
Researchers have developed a new in-sensor computing framework for energy-efficient Earth observation from satellites. This approach integrates TinyML techniques with the Sony IMX500 Intelligent Vision Sensor to process…
-
New JEPA Architectures Achieve Stable End-to-End Training from Pixels
Researchers have developed LeWorldModel (LeWM), a novel Joint Embedding Predictive Architecture (JEPA) that stably trains end-to-end from raw pixels. Unlike previous fragile JEPA methods, LeWM uses only two loss terms a…
-
Location encoders improve AI satellite image analysis
A new benchmark study explores how to best incorporate geographic location data into AI models for satellite image analysis. Researchers tested three methods—naive sin/cos, GeoCLIP, and SatCLIP—to encode latitude and lo…
-
Trust-SSL enhances aerial image self-supervised learning robustness to degradation
Researchers have developed Trust-SSL, a novel self-supervised learning strategy designed to improve the robustness of aerial image analysis. This method introduces a per-sample trust weight into the alignment objective,…