Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385647141> ?p ?o ?g. }
Showing items 1 to 97 of
97
with 100 items per page.
- W4385647141 endingPage "64" @default.
- W4385647141 startingPage "50" @default.
- W4385647141 abstract "In image classification, a deep neural network (DNN) that is trained on undistorted images constitutes an effective decision boundary. Unfortunately, this boundary does not support distorted images, such as noisy or blurry ones, leading to accuracy drop-off. As a simple approach for classifying distorted images as well as undistorted ones, previous methods have optimized the trained DNN again on both kinds of images. However, in these methods, the decision boundary may become overly complicated during optimization because there is no regularization of the decision boundary. Consequently, this decision boundary limits efficient optimization. In this paper, we study a simple yet effective decision boundary for distorted image classification through the use of a novel loss, called a neural activation pattern matching (NAPM) loss. The NAPM loss is based on recent findings that the decision boundary is a piecewise linear function, where each linear segment is constructed from a neural activation pattern in the DNN when an image is fed to it. The NAPM loss extracts the neural activation patterns when the distorted image and its undistorted version are fed to the DNN and then matches them with each other via the sigmoid cross-entropy. Therefore, it constrains the DNN to classify the distorted image and its undistorted version by the same linear segment. As a result, our loss accelerates efficient optimization by preventing the decision boundary from becoming overly complicated. Our experiments demonstrate that our loss increases the accuracy of the previous methods in all conditions evaluated." @default.
- W4385647141 created "2023-08-09" @default.
- W4385647141 creator A5006634833 @default.
- W4385647141 creator A5037810535 @default.
- W4385647141 creator A5060947560 @default.
- W4385647141 creator A5063110877 @default.
- W4385647141 creator A5064042196 @default.
- W4385647141 date "2023-10-01" @default.
- W4385647141 modified "2023-10-18" @default.
- W4385647141 title "Distorted image classification using neural activation pattern matching loss" @default.
- W4385647141 cites W1849277567 @default.
- W4385647141 cites W2062118960 @default.
- W4385647141 cites W2097117768 @default.
- W4385647141 cites W2101926813 @default.
- W4385647141 cites W2108598243 @default.
- W4385647141 cites W2120480077 @default.
- W4385647141 cites W2147800946 @default.
- W4385647141 cites W2155904486 @default.
- W4385647141 cites W2194775991 @default.
- W4385647141 cites W2518775244 @default.
- W4385647141 cites W2578937925 @default.
- W4385647141 cites W2612573399 @default.
- W4385647141 cites W2626739722 @default.
- W4385647141 cites W2804935296 @default.
- W4385647141 cites W2887876256 @default.
- W4385647141 cites W2913340405 @default.
- W4385647141 cites W2963980515 @default.
- W4385647141 cites W2964111476 @default.
- W4385647141 cites W2964753892 @default.
- W4385647141 cites W2979543786 @default.
- W4385647141 cites W2982157312 @default.
- W4385647141 cites W3102278250 @default.
- W4385647141 cites W3209011223 @default.
- W4385647141 cites W4312282202 @default.
- W4385647141 doi "https://doi.org/10.1016/j.neunet.2023.07.050" @default.
- W4385647141 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37625242" @default.
- W4385647141 hasPublicationYear "2023" @default.
- W4385647141 type Work @default.
- W4385647141 citedByCount "0" @default.
- W4385647141 crossrefType "journal-article" @default.
- W4385647141 hasAuthorship W4385647141A5006634833 @default.
- W4385647141 hasAuthorship W4385647141A5037810535 @default.
- W4385647141 hasAuthorship W4385647141A5060947560 @default.
- W4385647141 hasAuthorship W4385647141A5063110877 @default.
- W4385647141 hasAuthorship W4385647141A5064042196 @default.
- W4385647141 hasBestOaLocation W43856471411 @default.
- W4385647141 hasConcept C115961682 @default.
- W4385647141 hasConcept C12267149 @default.
- W4385647141 hasConcept C134306372 @default.
- W4385647141 hasConcept C153180895 @default.
- W4385647141 hasConcept C154945302 @default.
- W4385647141 hasConcept C17095337 @default.
- W4385647141 hasConcept C2524010 @default.
- W4385647141 hasConcept C33923547 @default.
- W4385647141 hasConcept C38365724 @default.
- W4385647141 hasConcept C41008148 @default.
- W4385647141 hasConcept C42023084 @default.
- W4385647141 hasConcept C50644808 @default.
- W4385647141 hasConcept C62354387 @default.
- W4385647141 hasConcept C75294576 @default.
- W4385647141 hasConcept C81388566 @default.
- W4385647141 hasConceptScore W4385647141C115961682 @default.
- W4385647141 hasConceptScore W4385647141C12267149 @default.
- W4385647141 hasConceptScore W4385647141C134306372 @default.
- W4385647141 hasConceptScore W4385647141C153180895 @default.
- W4385647141 hasConceptScore W4385647141C154945302 @default.
- W4385647141 hasConceptScore W4385647141C17095337 @default.
- W4385647141 hasConceptScore W4385647141C2524010 @default.
- W4385647141 hasConceptScore W4385647141C33923547 @default.
- W4385647141 hasConceptScore W4385647141C38365724 @default.
- W4385647141 hasConceptScore W4385647141C41008148 @default.
- W4385647141 hasConceptScore W4385647141C42023084 @default.
- W4385647141 hasConceptScore W4385647141C50644808 @default.
- W4385647141 hasConceptScore W4385647141C62354387 @default.
- W4385647141 hasConceptScore W4385647141C75294576 @default.
- W4385647141 hasConceptScore W4385647141C81388566 @default.
- W4385647141 hasFunder F4320334764 @default.
- W4385647141 hasLocation W43856471411 @default.
- W4385647141 hasLocation W43856471412 @default.
- W4385647141 hasOpenAccess W4385647141 @default.
- W4385647141 hasPrimaryLocation W43856471411 @default.
- W4385647141 hasRelatedWork W1504105233 @default.
- W4385647141 hasRelatedWork W2041004593 @default.
- W4385647141 hasRelatedWork W2811324119 @default.
- W4385647141 hasRelatedWork W2895192346 @default.
- W4385647141 hasRelatedWork W3024979424 @default.
- W4385647141 hasRelatedWork W3032499992 @default.
- W4385647141 hasRelatedWork W3110577345 @default.
- W4385647141 hasRelatedWork W3134817226 @default.
- W4385647141 hasRelatedWork W3136579697 @default.
- W4385647141 hasRelatedWork W4283785902 @default.
- W4385647141 hasVolume "167" @default.
- W4385647141 isParatext "false" @default.
- W4385647141 isRetracted "false" @default.
- W4385647141 workType "article" @default.