Deeplabv3 Plus
PulseAugur coverage of Deeplabv3 Plus — every cluster mentioning Deeplabv3 Plus across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
New AI framework offers accurate thyroid nodule segmentation with uncertainty estimates
Researchers have developed a new framework for segmenting thyroid nodules in ultrasound images, integrating DeepLabV3+ with channel-attention modules and post-hoc conformal prediction. This approach not only provides ac…
-
New deep learning patching method balances performance and training time
A new paper introduces Stride Independent Patching (SSP), a semi-automatic method for positioning image patches in deep learning. Unlike automatic stride-dependent methods, SSP relies on user input to define patch locat…
-
FCBNet offers parameter-efficient weed detection in aerial imagery
Researchers have developed FCBNet, a parameter-efficient convolutional neural network designed for detecting camouflaged weeds in multispectral aerial imagery. The model utilizes a frozen ConvNeXt backbone and a novel F…
-
New MeCSAFNet architecture improves multispectral land cover segmentation
Researchers have introduced MeCSAFNet, a novel multi-encoder architecture designed for semantic segmentation of multispectral imagery, specifically for land cover classification. The network utilizes dual ConvNeXt encod…
-
New Evidential Deep Learning method improves OOD detection in segmentation
Researchers have developed a novel approach to improve out-of-distribution (OOD) detection in semantic segmentation networks by leveraging Wasserstein-based objectives within an Evidential Deep Learning (EDL) framework.…
-
New diffusion model accurately detects aircraft contrails from satellite data
Researchers have developed a diffusion model capable of detecting contrails in satellite imagery with improved accuracy. The model, trained on a single GPU, achieved a PR-AUC of 0.476, significantly outperforming baseli…
-
New research compares AI vision models to human physical reasoning
Researchers have developed a new method to compare the object representations learned by computer vision models with those of humans. Using a time-to-collision and change detection task with 178 and 50 human participant…
-
New pipeline enhances teledermatology image quality assessment
Researchers have developed a new pipeline called the Semantic Tri-view Pipeline to improve the gradability of teledermatology images. This system analyzes skin micro-relief across multiple photographic views to assess d…
-
DWFF-Net advances multi-scale segmentation for agricultural habitat mapping
Researchers have developed DWFF-Net, a novel method for multi-scale segmentation in agricultural habitat identification. This network utilizes a frozen DINOv3 encoder for feature extraction and incorporates an adaptive …
-
New method improves wildfire segmentation using Landsat-8 imagery
Researchers have developed a new method for segmenting active wildfires using Landsat-8 satellite imagery, addressing the challenge of sparse and imbalanced fire pixel data. The study evaluated three segmentation archit…
-
UAV wildfire segmentation benefits from RGB-infrared fusion, study finds
Researchers have conducted a comparative study on multimodal RGB-infrared fusion for wildfire segmentation using unmanned aerial vehicles (UAVs). The study evaluated three fusion strategies across U-Net, DeepLabV3+, and…
-
New deep learning benchmark MeltwaterBench improves Greenland meltwater mapping
Researchers have developed MeltwaterBench, a new deep learning framework designed to improve the spatiotemporal resolution of surface meltwater maps from the Greenland ice sheet. This model fuses remote sensing data wit…
-
New research explores uncertainty quantification and lightweight models for semantic segmentation
Researchers are exploring methods to improve the reliability and robustness of semantic segmentation models, particularly for safety-critical applications. One paper investigates the integration of uncertainty quantific…
-
Deep learning models benchmarked for lung cancer histopathology analysis
Researchers have developed a two-stage deep learning framework for analyzing lung cancer histopathology images. The framework systematically compares state-of-the-art architectures for both tissue classification and reg…
-
New framework enhances plant stress phenotyping with diffusion-guided segmentation
Researchers have developed a novel diffusion-guided hybrid segmentation framework designed to improve the accuracy and efficiency of plant stress phenotyping in agricultural imagery. This framework combines established …
-
FSB-Net improves brain stroke lesion segmentation using frequency-spatial analysis
Researchers have developed FSB-Net, a novel deep learning model designed for precise segmentation of brain stroke lesions in non-contrast CT scans. This network uniquely incorporates frequency-domain analysis to better …
-
New framework improves building detection using Sentinel-2 satellite data
Researchers have developed a framework for robust building detection using Sentinel-2 satellite imagery, addressing challenges posed by the imagery's 10m resolution and variations in seasonality and urban environments. …
-
New architecture tackles diabetic retinopathy lesion segmentation challenges
Researchers have developed a new deep learning architecture called the Multi-Resolution Feature Stem to improve the segmentation of diabetic retinopathy lesions. Existing models struggle because DR lesions vary signific…
-
New framework enhances welding robot seam segmentation with transfer learning · 2 sources tracked
Researchers have developed a new framework to improve seam segmentation for automated welding robots in construction, addressing challenges like harsh lighting and reflections. The approach enhances the BiSeNetV2 model …
-
Transformer models show better generalization in diabetic foot ulcer segmentation
A new study benchmarks three deep learning models for diabetic foot ulcer segmentation: U-Net, DeepLabV3+, and SegFormer-B2. While all models performed well on their training datasets, their accuracy significantly degra…