Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386075995> ?p ?o ?g. }
- W4386075995 abstract "Existing semantic segmentation methods improve generalization capability, by regularizing various images to a canonical feature space. While this process contributes to generalization, it weakens the representation inevitably. In contrast to existing methods, we instead utilize the difference between images to build a better representation space, where the distinct style features are extracted and stored as the bases of representation. Then, the generalization to unseen image styles is achieved by projecting features to this known space. Specifically, we realize the style projection as a weighted combination of stored bases, where the similarity distances are adopted as the weighting factors. Based on the same concept, we extend this process to the decision part of model and promote the generalization of semantic prediction. By measuring the similarity distances to semantic bases (i.e., prototypes), we replace the common deterministic prediction with semantic clustering. Comprehensive experiments demonstrate the advantage of proposed method to the state of the art, up to 3.6% mIoU improvement in average on unseen scenarios. Code and models are available at https://gitee.com/mindspore/models/tree/master/research/cv/SPC-Net." @default.
- W4386075995 created "2023-08-23" @default.
- W4386075995 creator A5008612863 @default.
- W4386075995 creator A5012400248 @default.
- W4386075995 creator A5022598327 @default.
- W4386075995 creator A5033093547 @default.
- W4386075995 creator A5053491626 @default.
- W4386075995 creator A5081419302 @default.
- W4386075995 creator A5082226210 @default.
- W4386075995 creator A5091053091 @default.
- W4386075995 date "2023-06-01" @default.
- W4386075995 modified "2023-10-16" @default.
- W4386075995 title "Style Projected Clustering for Domain Generalized Semantic Segmentation" @default.
- W4386075995 cites W2078451758 @default.
- W4386075995 cites W2086234878 @default.
- W4386075995 cites W2108598243 @default.
- W4386075995 cites W2122111042 @default.
- W4386075995 cites W2150066425 @default.
- W4386075995 cites W2151537585 @default.
- W4386075995 cites W2194775991 @default.
- W4386075995 cites W2340897893 @default.
- W4386075995 cites W2431874326 @default.
- W4386075995 cites W2487365028 @default.
- W4386075995 cites W2572730214 @default.
- W4386075995 cites W2603777577 @default.
- W4386075995 cites W2763549966 @default.
- W4386075995 cites W2781228439 @default.
- W4386075995 cites W2798658180 @default.
- W4386075995 cites W2798991696 @default.
- W4386075995 cites W2883780447 @default.
- W4386075995 cites W2884366600 @default.
- W4386075995 cites W2894728917 @default.
- W4386075995 cites W2963043696 @default.
- W4386075995 cites W2963073217 @default.
- W4386075995 cites W2963107255 @default.
- W4386075995 cites W2963163009 @default.
- W4386075995 cites W2963403405 @default.
- W4386075995 cites W2963743626 @default.
- W4386075995 cites W2981488580 @default.
- W4386075995 cites W2981540341 @default.
- W4386075995 cites W2990230185 @default.
- W4386075995 cites W2995182785 @default.
- W4386075995 cites W3034562924 @default.
- W4386075995 cites W3034730995 @default.
- W4386075995 cites W3035564946 @default.
- W4386075995 cites W3092612036 @default.
- W4386075995 cites W3099206234 @default.
- W4386075995 cites W3108560336 @default.
- W4386075995 cites W3122412340 @default.
- W4386075995 cites W3133542152 @default.
- W4386075995 cites W3166542014 @default.
- W4386075995 cites W3169545167 @default.
- W4386075995 cites W3172292202 @default.
- W4386075995 cites W3172863135 @default.
- W4386075995 cites W3175308890 @default.
- W4386075995 cites W3180003570 @default.
- W4386075995 cites W3180439858 @default.
- W4386075995 cites W3181515393 @default.
- W4386075995 cites W4205474609 @default.
- W4386075995 cites W4221166019 @default.
- W4386075995 cites W4226040545 @default.
- W4386075995 cites W4289639938 @default.
- W4386075995 cites W4312657985 @default.
- W4386075995 cites W4312719508 @default.
- W4386075995 cites W4313011609 @default.
- W4386075995 cites W4313023122 @default.
- W4386075995 cites W4313045831 @default.
- W4386075995 doi "https://doi.org/10.1109/cvpr52729.2023.00299" @default.
- W4386075995 hasPublicationYear "2023" @default.
- W4386075995 type Work @default.
- W4386075995 citedByCount "0" @default.
- W4386075995 crossrefType "proceedings-article" @default.
- W4386075995 hasAuthorship W4386075995A5008612863 @default.
- W4386075995 hasAuthorship W4386075995A5012400248 @default.
- W4386075995 hasAuthorship W4386075995A5022598327 @default.
- W4386075995 hasAuthorship W4386075995A5033093547 @default.
- W4386075995 hasAuthorship W4386075995A5053491626 @default.
- W4386075995 hasAuthorship W4386075995A5081419302 @default.
- W4386075995 hasAuthorship W4386075995A5082226210 @default.
- W4386075995 hasAuthorship W4386075995A5091053091 @default.
- W4386075995 hasConcept C103278499 @default.
- W4386075995 hasConcept C11413529 @default.
- W4386075995 hasConcept C115961682 @default.
- W4386075995 hasConcept C126838900 @default.
- W4386075995 hasConcept C130318100 @default.
- W4386075995 hasConcept C134306372 @default.
- W4386075995 hasConcept C138885662 @default.
- W4386075995 hasConcept C153180895 @default.
- W4386075995 hasConcept C154945302 @default.
- W4386075995 hasConcept C177148314 @default.
- W4386075995 hasConcept C17744445 @default.
- W4386075995 hasConcept C183115368 @default.
- W4386075995 hasConcept C199539241 @default.
- W4386075995 hasConcept C2776359362 @default.
- W4386075995 hasConcept C2776401178 @default.
- W4386075995 hasConcept C33923547 @default.
- W4386075995 hasConcept C41008148 @default.
- W4386075995 hasConcept C41895202 @default.
- W4386075995 hasConcept C57493831 @default.
- W4386075995 hasConcept C71924100 @default.