Matches in SemOpenAlex for { <https://semopenalex.org/work/W4316039456> ?p ?o ?g. }
- W4316039456 endingPage "215" @default.
- W4316039456 startingPage "202" @default.
- W4316039456 abstract "Nowadays many semantic segmentation algorithms have achieved satisfactory accuracy on von Neumann platforms (e.g., GPU), but the speed and energy consumption have not meet the high requirements of certain edge applications like autonomous driving. To tackle this issue, it is of necessity to design an efficient lightweight semantic segmentation algorithm and then implement it on emerging hardware platforms with high speed and energy efficiency. Here, we first propose an extremely factorized network (EFNet) which can learn multi-scale context information while preserving rich spatial information with reduced model complexity. Experimental results on the Cityscapes dataset show that EFNet achieves an accuracy of 68.0% mean intersection over union (mIoU) with only 0.18M parameters, at a speed of 99 frames per second (FPS) on a single RTX 3090 GPU. Then, to further improve the speed and energy efficiency, we design a memristor-based computing-in-memory (CIM) accelerator for the hardware implementation of EFNet. It is shown by the simulation in DNN+NeuroSim V2.0 that the memristor-based CIM accelerator is ∼63× (∼4.6×) smaller in area, at most ∼9.2× (∼1000×) faster, and ∼470× (∼2400×) more energy-efficient than the RTX 3090 GPU (the Jetson Nano embedded development board), although its accuracy slightly decreases by 1.7% mIoU. Therefore, the memristor-based CIM accelerator has great potential to be deployed at the edge to implement lightweight semantic segmentation models like EFNet. This study showcases an algorithm-hardware co-design to realize real-time and low-power semantic segmentation at the edge." @default.
- W4316039456 created "2023-01-14" @default.
- W4316039456 creator A5000817573 @default.
- W4316039456 creator A5007937560 @default.
- W4316039456 creator A5021736861 @default.
- W4316039456 creator A5025127998 @default.
- W4316039456 creator A5026877740 @default.
- W4316039456 creator A5034557983 @default.
- W4316039456 creator A5034752899 @default.
- W4316039456 creator A5038288450 @default.
- W4316039456 creator A5059075050 @default.
- W4316039456 creator A5064263084 @default.
- W4316039456 date "2023-03-01" @default.
- W4316039456 modified "2023-10-16" @default.
- W4316039456 title "Performance estimation for the memristor-based computing-in-memory implementation of extremely factorized network for real-time and low-power semantic segmentation" @default.
- W4316039456 cites W1969067576 @default.
- W4316039456 cites W2063695057 @default.
- W4316039456 cites W2171943915 @default.
- W4316039456 cites W2195569869 @default.
- W4316039456 cites W2397312330 @default.
- W4316039456 cites W2412782625 @default.
- W4316039456 cites W2442974303 @default.
- W4316039456 cites W2508602506 @default.
- W4316039456 cites W2529441728 @default.
- W4316039456 cites W2762439315 @default.
- W4316039456 cites W2782046614 @default.
- W4316039456 cites W2793132020 @default.
- W4316039456 cites W2796472956 @default.
- W4316039456 cites W2805362231 @default.
- W4316039456 cites W2823072336 @default.
- W4316039456 cites W2894293853 @default.
- W4316039456 cites W2928133111 @default.
- W4316039456 cites W2943925420 @default.
- W4316039456 cites W2953901595 @default.
- W4316039456 cites W2963881378 @default.
- W4316039456 cites W2966550174 @default.
- W4316039456 cites W3003821665 @default.
- W4316039456 cites W3007703297 @default.
- W4316039456 cites W3015379147 @default.
- W4316039456 cites W3026002147 @default.
- W4316039456 cites W3112503277 @default.
- W4316039456 cites W3112740243 @default.
- W4316039456 cites W3119749347 @default.
- W4316039456 cites W3128892380 @default.
- W4316039456 cites W3134439218 @default.
- W4316039456 cites W3145782339 @default.
- W4316039456 cites W3158385340 @default.
- W4316039456 cites W3159420753 @default.
- W4316039456 cites W3159640305 @default.
- W4316039456 cites W3159837173 @default.
- W4316039456 cites W3195338688 @default.
- W4316039456 cites W4281874271 @default.
- W4316039456 doi "https://doi.org/10.1016/j.neunet.2023.01.008" @default.
- W4316039456 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/36657333" @default.
- W4316039456 hasPublicationYear "2023" @default.
- W4316039456 type Work @default.
- W4316039456 citedByCount "0" @default.
- W4316039456 crossrefType "journal-article" @default.
- W4316039456 hasAuthorship W4316039456A5000817573 @default.
- W4316039456 hasAuthorship W4316039456A5007937560 @default.
- W4316039456 hasAuthorship W4316039456A5021736861 @default.
- W4316039456 hasAuthorship W4316039456A5025127998 @default.
- W4316039456 hasAuthorship W4316039456A5026877740 @default.
- W4316039456 hasAuthorship W4316039456A5034557983 @default.
- W4316039456 hasAuthorship W4316039456A5034752899 @default.
- W4316039456 hasAuthorship W4316039456A5038288450 @default.
- W4316039456 hasAuthorship W4316039456A5059075050 @default.
- W4316039456 hasAuthorship W4316039456A5064263084 @default.
- W4316039456 hasConcept C105795698 @default.
- W4316039456 hasConcept C111919701 @default.
- W4316039456 hasConcept C113775141 @default.
- W4316039456 hasConcept C119599485 @default.
- W4316039456 hasConcept C127413603 @default.
- W4316039456 hasConcept C146978453 @default.
- W4316039456 hasConcept C150072547 @default.
- W4316039456 hasConcept C151730666 @default.
- W4316039456 hasConcept C154945302 @default.
- W4316039456 hasConcept C162307627 @default.
- W4316039456 hasConcept C186370098 @default.
- W4316039456 hasConcept C24326235 @default.
- W4316039456 hasConcept C2742236 @default.
- W4316039456 hasConcept C2779343474 @default.
- W4316039456 hasConcept C2780165032 @default.
- W4316039456 hasConcept C3261483 @default.
- W4316039456 hasConcept C33923547 @default.
- W4316039456 hasConcept C41008148 @default.
- W4316039456 hasConcept C64543145 @default.
- W4316039456 hasConcept C79403827 @default.
- W4316039456 hasConcept C80469333 @default.
- W4316039456 hasConcept C86803240 @default.
- W4316039456 hasConcept C89600930 @default.
- W4316039456 hasConcept C9390403 @default.
- W4316039456 hasConceptScore W4316039456C105795698 @default.
- W4316039456 hasConceptScore W4316039456C111919701 @default.
- W4316039456 hasConceptScore W4316039456C113775141 @default.
- W4316039456 hasConceptScore W4316039456C119599485 @default.
- W4316039456 hasConceptScore W4316039456C127413603 @default.
- W4316039456 hasConceptScore W4316039456C146978453 @default.