This survey and benchmark of deep time series models using the released TSLib library finds that models with specific structures perform well only on distinct analysis tasks.
Deep residual learning for image recognition,
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 7roles
background 1polarities
background 1representative citing papers
Switchable Normalization learns per-layer weights to combine channel, layer, and minibatch normalizers, claiming robustness to batch size and better results than fixed normalizers on ImageNet, COCO, CityScapes, ADE20K, MegaFace, and Kinetics.
Adversarial perturbations disrupt DNN-based face detectors under white-box, gray-box, and black-box settings to sabotage training data for AI face synthesis.
Cross Attention Network fuses spatial and contextual features via a cross attention module to improve semantic segmentation performance and speed on Cityscapes and CamVid.
Kinematic JIPDA ranked first on 3DMOT2015 for fixed cameras while global nearest-neighbor on deep embeddings performed best for moving cameras; mixing embeddings into JIPDA added little.
DRLN uses residual-on-residual cascading, dense block concatenation, and Laplacian attention to learn multi-scale features for super-resolution with claimed favorable results on standard benchmarks.
The paper reports a multi-stage system for activity detection in extended videos that uses spatial object detections, temporal localization, tubelet generation variants, and late fusion of component outputs.
citing papers explorer
-
Deep Time Series Models: A Comprehensive Survey and Benchmark
This survey and benchmark of deep time series models using the released TSLib library finds that models with specific structures perform well only on distinct analysis tasks.
-
Switchable Normalization for Learning-to-Normalize Deep Representation
Switchable Normalization learns per-layer weights to combine channel, layer, and minibatch normalizers, claiming robustness to batch size and better results than fixed normalizers on ImageNet, COCO, CityScapes, ADE20K, MegaFace, and Kinetics.
-
Hiding Faces in Plain Sight: Disrupting AI Face Synthesis with Adversarial Perturbations
Adversarial perturbations disrupt DNN-based face detectors under white-box, gray-box, and black-box settings to sabotage training data for AI face synthesis.
-
Cross Attention Network for Semantic Segmentation
Cross Attention Network fuses spatial and contextual features via a cross attention module to improve semantic segmentation performance and speed on Cityscapes and CamVid.
-
Pedestrian Tracking by Probabilistic Data Association and Correspondence Embeddings
Kinematic JIPDA ranked first on 3DMOT2015 for fixed cameras while global nearest-neighbor on deep embeddings performed best for moving cameras; mixing embeddings into JIPDA added little.
-
Densely Residual Laplacian Super-Resolution
DRLN uses residual-on-residual cascading, dense block concatenation, and Laplacian attention to learn multi-scale features for super-resolution with claimed favorable results on standard benchmarks.
-
vireoJD-MM at Activity Detection in Extended Videos
The paper reports a multi-stage system for activity detection in extended videos that uses spatial object detections, temporal localization, tubelet generation variants, and late fusion of component outputs.