In the sub-wavelength regime,design for manufacturability（DFM） becomes increasingly important for field programmable gate arrays（FPGAs）.In this paper,an automated tile generation flow targeting micro-regular fabri...In the sub-wavelength regime,design for manufacturability（DFM） becomes increasingly important for field programmable gate arrays（FPGAs）.In this paper,an automated tile generation flow targeting micro-regular fabric is reported.Using a publicly accessible,well-documented academic FPGA as a case study,we found that compared to the tile generators previously reported,our generated micro-regular tile incurs less than 10%area overhead,which could be potentially recovered by process window optimization,thanks to its superior printability. In addition,we demonstrate that on 45 nm technology,the generated FPGA tile reduces lithography induced process variation by 33%,and reduce probability of failure by 21.2%.If a further overhead of 10%area can be recovered by enhanced resolution,we can achieve the variation reduction of 93.8%and reduce the probability of failure by 16.2%.展开更多
Facade parsing aims to decompose a building facade image into semantic regions of the facade objects.Considering each architectural element on a facade as a parameterized rectangle,we formulate the facade parsing task...Facade parsing aims to decompose a building facade image into semantic regions of the facade objects.Considering each architectural element on a facade as a parameterized rectangle,we formulate the facade parsing task as object detection,allowing overlapping and nesting,which will support structural 3D modeling and editing for further applications.In contrast to general object detection,the spatial arrangement regularity and appearance similarity between the facade elements of the same category provide valuable context for accurate element localization.In this paper,we propose to exploit the spatial arrangement regularity and appearance similarity of facade elements in a detection framework.Our element-arrangement context network(EACNet)consists of two unidirectional attention branches,one to capture the column-context and the other to capture row-context to aggregate element-specific features from multiple instances on the facade.We conduct extensive experiments on four public datasets(ECP,CMP,Graz50,and eTRIMS).The proposed EACNet achieves the highest mIoU(82.1%on ECP,77.35%on Graz50,and 82.3%on eTRIMS)compared with the state-of-the-art methods.Both the quantitative and qualitative evaluation results demonstrate the effectiveness of our dual unidirectional attention branches to parse facade elements.展开更多
文摘In the sub-wavelength regime,design for manufacturability（DFM） becomes increasingly important for field programmable gate arrays（FPGAs）.In this paper,an automated tile generation flow targeting micro-regular fabric is reported.Using a publicly accessible,well-documented academic FPGA as a case study,we found that compared to the tile generators previously reported,our generated micro-regular tile incurs less than 10%area overhead,which could be potentially recovered by process window optimization,thanks to its superior printability. In addition,we demonstrate that on 45 nm technology,the generated FPGA tile reduces lithography induced process variation by 33%,and reduce probability of failure by 21.2%.If a further overhead of 10%area can be recovered by enhanced resolution,we can achieve the variation reduction of 93.8%and reduce the probability of failure by 16.2%.
基金the National Natural Science Foundation of China under Grant No.61632006Tencent Corporation。
文摘Facade parsing aims to decompose a building facade image into semantic regions of the facade objects.Considering each architectural element on a facade as a parameterized rectangle,we formulate the facade parsing task as object detection,allowing overlapping and nesting,which will support structural 3D modeling and editing for further applications.In contrast to general object detection,the spatial arrangement regularity and appearance similarity between the facade elements of the same category provide valuable context for accurate element localization.In this paper,we propose to exploit the spatial arrangement regularity and appearance similarity of facade elements in a detection framework.Our element-arrangement context network(EACNet)consists of two unidirectional attention branches,one to capture the column-context and the other to capture row-context to aggregate element-specific features from multiple instances on the facade.We conduct extensive experiments on four public datasets(ECP,CMP,Graz50,and eTRIMS).The proposed EACNet achieves the highest mIoU(82.1%on ECP,77.35%on Graz50,and 82.3%on eTRIMS)compared with the state-of-the-art methods.Both the quantitative and qualitative evaluation results demonstrate the effectiveness of our dual unidirectional attention branches to parse facade elements.
