Matches in SemOpenAlex for { <https://semopenalex.org/work/W4220932940> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W4220932940 abstract "Optimization of CT image quality typically involves balancing noise and bias. In filtered back-projection, this trade-off is controlled by the particular filter and cutoff frequency. In penalized-likelihood iterative reconstruction, the penalty weight serves the same function. Deep neural networks typically do not provide this tuneable control over output image properties. Models are often trained to minimize mean squared error which penalizes both variance and bias in image outputs but does not offer any control over the trade-off between the two. In this work, we propose a method for controlling the output image properties of neural networks with a new loss function called weighted covariance and bias (WCB). Our proposed method includes separate weighting parameters to control the relative importance of noise or bias reduction. Moreover, we show that tuning these weights enables targeted penalization of specific image features (e.g. spatial frequencies). To evaluate our method, we present a simulation study using digital anthropormorphic phantoms, physical simulation of non-ideal CT data, and image formation with various algorithms. We show that WCB offers a greater degree of control over trade-offs between variance and bias whereas MSE has only one configuration. We also show that WCB can be used to control specific image properties including variance, bias, spatial resolution, and the noise correlation of neural network outputs. Finally, we present a method to optimize the proposed weights for stimulus detectability. Our results demonstrate the potential for this new capability to control the image properties of DNN outputs and optimize image quality for the task-specific applications." @default.
- W4220932940 created "2022-04-03" @default.
- W4220932940 creator A5001358171 @default.
- W4220932940 creator A5023257111 @default.
- W4220932940 creator A5050134612 @default.
- W4220932940 creator A5072931395 @default.
- W4220932940 creator A5089321471 @default.
- W4220932940 date "2022-03-31" @default.
- W4220932940 modified "2023-09-30" @default.
- W4220932940 title "Control of variance and bias in CT image processing with variational training of deep neural networks" @default.
- W4220932940 cites W1983281817 @default.
- W4220932940 cites W2071847032 @default.
- W4220932940 cites W2141689871 @default.
- W4220932940 cites W2766371297 @default.
- W4220932940 cites W3101078346 @default.
- W4220932940 cites W3169967744 @default.
- W4220932940 doi "https://doi.org/10.1117/12.2612417" @default.
- W4220932940 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/35656120" @default.
- W4220932940 hasPublicationYear "2022" @default.
- W4220932940 type Work @default.
- W4220932940 citedByCount "2" @default.
- W4220932940 countsByYear W42209329402022 @default.
- W4220932940 countsByYear W42209329402023 @default.
- W4220932940 crossrefType "proceedings-article" @default.
- W4220932940 hasAuthorship W4220932940A5001358171 @default.
- W4220932940 hasAuthorship W4220932940A5023257111 @default.
- W4220932940 hasAuthorship W4220932940A5050134612 @default.
- W4220932940 hasAuthorship W4220932940A5072931395 @default.
- W4220932940 hasAuthorship W4220932940A5089321471 @default.
- W4220932940 hasBestOaLocation W42209329402 @default.
- W4220932940 hasConcept C105795698 @default.
- W4220932940 hasConcept C11413529 @default.
- W4220932940 hasConcept C115961682 @default.
- W4220932940 hasConcept C126838900 @default.
- W4220932940 hasConcept C139945424 @default.
- W4220932940 hasConcept C141379421 @default.
- W4220932940 hasConcept C153180895 @default.
- W4220932940 hasConcept C154945302 @default.
- W4220932940 hasConcept C183115368 @default.
- W4220932940 hasConcept C33923547 @default.
- W4220932940 hasConcept C41008148 @default.
- W4220932940 hasConcept C50644808 @default.
- W4220932940 hasConcept C55020928 @default.
- W4220932940 hasConcept C71924100 @default.
- W4220932940 hasConcept C99498987 @default.
- W4220932940 hasConceptScore W4220932940C105795698 @default.
- W4220932940 hasConceptScore W4220932940C11413529 @default.
- W4220932940 hasConceptScore W4220932940C115961682 @default.
- W4220932940 hasConceptScore W4220932940C126838900 @default.
- W4220932940 hasConceptScore W4220932940C139945424 @default.
- W4220932940 hasConceptScore W4220932940C141379421 @default.
- W4220932940 hasConceptScore W4220932940C153180895 @default.
- W4220932940 hasConceptScore W4220932940C154945302 @default.
- W4220932940 hasConceptScore W4220932940C183115368 @default.
- W4220932940 hasConceptScore W4220932940C33923547 @default.
- W4220932940 hasConceptScore W4220932940C41008148 @default.
- W4220932940 hasConceptScore W4220932940C50644808 @default.
- W4220932940 hasConceptScore W4220932940C55020928 @default.
- W4220932940 hasConceptScore W4220932940C71924100 @default.
- W4220932940 hasConceptScore W4220932940C99498987 @default.
- W4220932940 hasLocation W42209329401 @default.
- W4220932940 hasLocation W42209329402 @default.
- W4220932940 hasLocation W42209329403 @default.
- W4220932940 hasOpenAccess W4220932940 @default.
- W4220932940 hasPrimaryLocation W42209329401 @default.
- W4220932940 hasRelatedWork W1965232618 @default.
- W4220932940 hasRelatedWork W2028581306 @default.
- W4220932940 hasRelatedWork W2031845464 @default.
- W4220932940 hasRelatedWork W2043010645 @default.
- W4220932940 hasRelatedWork W2116558684 @default.
- W4220932940 hasRelatedWork W2127592224 @default.
- W4220932940 hasRelatedWork W2165186741 @default.
- W4220932940 hasRelatedWork W3176888188 @default.
- W4220932940 hasRelatedWork W4319603062 @default.
- W4220932940 hasRelatedWork W4362693467 @default.
- W4220932940 isParatext "false" @default.
- W4220932940 isRetracted "false" @default.
- W4220932940 workType "article" @default.