Matches in SemOpenAlex for { <https://semopenalex.org/work/W3009950151> ?p ?o ?g. }
Showing items 1 to 53 of
53
with 100 items per page.
- W3009950151 abstract "Semantic segmentation is the classification of each pixel in an image to an object, the resultant pixel map has significant usage in many fields. Some fields where this technology is being actively researched is in medicine, agriculture and robotics. For uses where the resources or power requirements are restricted such as robotics or where large amounts of images are required to process, efficiency can be key to the feasibility of a technique. Other applications that require real-time processing have a need for fast and efficient methods, especially where collision avoidance or safety may be involved. We take a combination of existing semantic segmentation methods and improve upon the efficiency by the replacement of the decoder network in ERFNet with a method based upon Dense Upscaling Convolutions, we then add a novel layer that allows the fine tuning of the decoder channel depth and therefore the efficiency of the network. Our proposed modification achieves 20-30% improvement in efficiency on moderate hardware (Nvidia GTX 960) over the original ERFNET and an additional 10% efficiency over the original Dense Upscaling Convolution. We perform a series of experiments to determine viable hyperparameters for the modification and measure the efficiency and accuracy over a range of image sizes, proving the viability of our approach." @default.
- W3009950151 created "2020-03-13" @default.
- W3009950151 creator A5041568282 @default.
- W3009950151 creator A5090503795 @default.
- W3009950151 creator A5091642155 @default.
- W3009950151 date "2020-01-12" @default.
- W3009950151 modified "2023-09-25" @default.
- W3009950151 title "Efficient Semantic Segmentation through Dense Upscaling Convolutions" @default.
- W3009950151 cites W1903029394 @default.
- W3009950151 cites W2412782625 @default.
- W3009950151 cites W2476548250 @default.
- W3009950151 cites W2762439315 @default.
- W3009950151 cites W2804860796 @default.
- W3009950151 doi "https://doi.org/10.1145/3378936.3378941" @default.
- W3009950151 hasPublicationYear "2020" @default.
- W3009950151 type Work @default.
- W3009950151 sameAs 3009950151 @default.
- W3009950151 citedByCount "0" @default.
- W3009950151 crossrefType "proceedings-article" @default.
- W3009950151 hasAuthorship W3009950151A5041568282 @default.
- W3009950151 hasAuthorship W3009950151A5090503795 @default.
- W3009950151 hasAuthorship W3009950151A5091642155 @default.
- W3009950151 hasConcept C124504099 @default.
- W3009950151 hasConcept C154945302 @default.
- W3009950151 hasConcept C184337299 @default.
- W3009950151 hasConcept C199360897 @default.
- W3009950151 hasConcept C204321447 @default.
- W3009950151 hasConcept C41008148 @default.
- W3009950151 hasConcept C89600930 @default.
- W3009950151 hasConceptScore W3009950151C124504099 @default.
- W3009950151 hasConceptScore W3009950151C154945302 @default.
- W3009950151 hasConceptScore W3009950151C184337299 @default.
- W3009950151 hasConceptScore W3009950151C199360897 @default.
- W3009950151 hasConceptScore W3009950151C204321447 @default.
- W3009950151 hasConceptScore W3009950151C41008148 @default.
- W3009950151 hasConceptScore W3009950151C89600930 @default.
- W3009950151 hasLocation W30099501511 @default.
- W3009950151 hasOpenAccess W3009950151 @default.
- W3009950151 hasPrimaryLocation W30099501511 @default.
- W3009950151 hasRelatedWork W2045391057 @default.
- W3009950151 hasRelatedWork W2119108994 @default.
- W3009950151 hasRelatedWork W2149145677 @default.
- W3009950151 hasRelatedWork W2152591411 @default.
- W3009950151 hasRelatedWork W2548218638 @default.
- W3009950151 hasRelatedWork W2551390060 @default.
- W3009950151 hasRelatedWork W2948522034 @default.
- W3009950151 hasRelatedWork W3107474891 @default.
- W3009950151 hasRelatedWork W3127804355 @default.
- W3009950151 hasRelatedWork W3161321444 @default.
- W3009950151 isParatext "false" @default.
- W3009950151 isRetracted "false" @default.
- W3009950151 magId "3009950151" @default.
- W3009950151 workType "article" @default.