Matches in SemOpenAlex for { <https://semopenalex.org/work/W4328011741> ?p ?o ?g. }
Showing items 1 to 80 of
80
with 100 items per page.
- W4328011741 abstract "Recently, 3D deep neural networks have been fully developed and applied to many high-safety tasks. However, due to the uninterpretability of deep learning networks, adversarial examples can easily prompt a normally trained deep learning model to make wrong predictions. In this paper, we propose a new point cloud defense network named DDR-Defense, a framework for defending neural network classifiers against adversarial examples. DDR-Defense neither modifies the number of the points in the input samples nor the protected classifiers so that it can protect most classification models. DDR-Defense first distinguishes adversarial examples from normal examples through a reconstruction-based detector. The detector can prevent errors caused by processing the entire input samples, thereby improving the security of the defense network. For adversarial examples, we first use the statistical outlier removal (SOR) method for denoising, then use a reformer to rebuild them. In this paper, We design a new reformer based on FoldingNet and variational autoencoder, named Folding-VAE. We test DDR-Defense on the ModelNet40 dataset and find that it has a better defense effect than other existing 3D defense networks, especially in saliency maps attack and LG-GAN attack. The lightweight detector, denoiser, and reformer framework ensures the security and efficiency of 3D defense for most application scenarios. Our research will provide a basis for improving the robustness of deep learning models on 3D point clouds." @default.
- W4328011741 created "2023-03-22" @default.
- W4328011741 creator A5003953815 @default.
- W4328011741 creator A5029959318 @default.
- W4328011741 creator A5032239710 @default.
- W4328011741 date "2022-12-09" @default.
- W4328011741 modified "2023-09-27" @default.
- W4328011741 title "DDR-Defense: 3D Defense Network with a Detector, a Denoiser, and a Reformer" @default.
- W4328011741 cites W2618043096 @default.
- W4328011741 cites W2796426482 @default.
- W4328011741 cites W2962818872 @default.
- W4328011741 cites W2971089407 @default.
- W4328011741 cites W2979750740 @default.
- W4328011741 cites W2981979099 @default.
- W4328011741 cites W2982104318 @default.
- W4328011741 cites W2990258745 @default.
- W4328011741 cites W2997532515 @default.
- W4328011741 cites W3008415104 @default.
- W4328011741 cites W3034376720 @default.
- W4328011741 cites W3170794970 @default.
- W4328011741 cites W3194885736 @default.
- W4328011741 cites W4214812658 @default.
- W4328011741 cites W4245534220 @default.
- W4328011741 doi "https://doi.org/10.1109/iccc56324.2022.10065933" @default.
- W4328011741 hasPublicationYear "2022" @default.
- W4328011741 type Work @default.
- W4328011741 citedByCount "0" @default.
- W4328011741 crossrefType "proceedings-article" @default.
- W4328011741 hasAuthorship W4328011741A5003953815 @default.
- W4328011741 hasAuthorship W4328011741A5029959318 @default.
- W4328011741 hasAuthorship W4328011741A5032239710 @default.
- W4328011741 hasConcept C101738243 @default.
- W4328011741 hasConcept C104317684 @default.
- W4328011741 hasConcept C108583219 @default.
- W4328011741 hasConcept C119857082 @default.
- W4328011741 hasConcept C154945302 @default.
- W4328011741 hasConcept C185592680 @default.
- W4328011741 hasConcept C2984842247 @default.
- W4328011741 hasConcept C37736160 @default.
- W4328011741 hasConcept C38652104 @default.
- W4328011741 hasConcept C41008148 @default.
- W4328011741 hasConcept C50644808 @default.
- W4328011741 hasConcept C55493867 @default.
- W4328011741 hasConcept C63479239 @default.
- W4328011741 hasConcept C739882 @default.
- W4328011741 hasConcept C76155785 @default.
- W4328011741 hasConcept C94915269 @default.
- W4328011741 hasConceptScore W4328011741C101738243 @default.
- W4328011741 hasConceptScore W4328011741C104317684 @default.
- W4328011741 hasConceptScore W4328011741C108583219 @default.
- W4328011741 hasConceptScore W4328011741C119857082 @default.
- W4328011741 hasConceptScore W4328011741C154945302 @default.
- W4328011741 hasConceptScore W4328011741C185592680 @default.
- W4328011741 hasConceptScore W4328011741C2984842247 @default.
- W4328011741 hasConceptScore W4328011741C37736160 @default.
- W4328011741 hasConceptScore W4328011741C38652104 @default.
- W4328011741 hasConceptScore W4328011741C41008148 @default.
- W4328011741 hasConceptScore W4328011741C50644808 @default.
- W4328011741 hasConceptScore W4328011741C55493867 @default.
- W4328011741 hasConceptScore W4328011741C63479239 @default.
- W4328011741 hasConceptScore W4328011741C739882 @default.
- W4328011741 hasConceptScore W4328011741C76155785 @default.
- W4328011741 hasConceptScore W4328011741C94915269 @default.
- W4328011741 hasFunder F4320321001 @default.
- W4328011741 hasLocation W43280117411 @default.
- W4328011741 hasOpenAccess W4328011741 @default.
- W4328011741 hasPrimaryLocation W43280117411 @default.
- W4328011741 hasRelatedWork W2669956259 @default.
- W4328011741 hasRelatedWork W2952919291 @default.
- W4328011741 hasRelatedWork W3044458868 @default.
- W4328011741 hasRelatedWork W3193857078 @default.
- W4328011741 hasRelatedWork W3208304128 @default.
- W4328011741 hasRelatedWork W3208723233 @default.
- W4328011741 hasRelatedWork W4213225422 @default.
- W4328011741 hasRelatedWork W4293054861 @default.
- W4328011741 hasRelatedWork W4311734044 @default.
- W4328011741 hasRelatedWork W4317552138 @default.
- W4328011741 isParatext "false" @default.
- W4328011741 isRetracted "false" @default.
- W4328011741 workType "article" @default.