In this paper, we reconsider the clustering problem for image over-segmentation from a new perspective. We propose a novel search algorithm called"active search" which explicitly considers neighbor continuit...In this paper, we reconsider the clustering problem for image over-segmentation from a new perspective. We propose a novel search algorithm called"active search" which explicitly considers neighbor continuity. Based on this search method, we design a back-and-forth traversal strategy and a joint assignment and update step to speed up the algorithm. Compared to earlier methods, such as simple linear iterative clustering(SLIC) and its variants, which use fixed search regions and perform the assignment and the update steps separately, our novel scheme reduces the number of iterations required for convergence,and also provides better boundaries in the oversegmentation results. Extensive evaluation using the Berkeley segmentation benchmark verifies that our method outperforms competing methods under various evaluation metrics. In particular, our method is fastest,achieving approximately 30 fps for a 481 × 321 image on a single CPU core. To facilitate further research, our code is made publicly available.展开更多
This paper provides a concrete and simple introduction to two pillars of domain theory : (1) solving recursive domain equations, and (2) universal and saturated domains. Our exposition combines Larsen and Winskel'...This paper provides a concrete and simple introduction to two pillars of domain theory : (1) solving recursive domain equations, and (2) universal and saturated domains. Our exposition combines Larsen and Winskel's idea on solving domain equations using information systems with Girard's idea of stable domain theory in the form of coherence spacest or graphs. Detailed constructions are given for universal and even homogeneous objects in two categories of graphs: one representing binary complete, prime algebraic domains with complete primes covering the bottom; the other representing w-algebraic, prime algebraic lattices. The back- and-forth argument in model theory helps to enlighten the constructions.展开更多
基金sponsored by National Natural Science Foundation of China (Nos. 61620106008 and 61572264)Huawei Innovation Research Program (HIRP)
文摘In this paper, we reconsider the clustering problem for image over-segmentation from a new perspective. We propose a novel search algorithm called"active search" which explicitly considers neighbor continuity. Based on this search method, we design a back-and-forth traversal strategy and a joint assignment and update step to speed up the algorithm. Compared to earlier methods, such as simple linear iterative clustering(SLIC) and its variants, which use fixed search regions and perform the assignment and the update steps separately, our novel scheme reduces the number of iterations required for convergence,and also provides better boundaries in the oversegmentation results. Extensive evaluation using the Berkeley segmentation benchmark verifies that our method outperforms competing methods under various evaluation metrics. In particular, our method is fastest,achieving approximately 30 fps for a 481 × 321 image on a single CPU core. To facilitate further research, our code is made publicly available.
基金This work is supported by the National Natural Science Foundation of China (No.69873034), the Foundation forUniversity Key Tea
文摘This paper provides a concrete and simple introduction to two pillars of domain theory : (1) solving recursive domain equations, and (2) universal and saturated domains. Our exposition combines Larsen and Winskel's idea on solving domain equations using information systems with Girard's idea of stable domain theory in the form of coherence spacest or graphs. Detailed constructions are given for universal and even homogeneous objects in two categories of graphs: one representing binary complete, prime algebraic domains with complete primes covering the bottom; the other representing w-algebraic, prime algebraic lattices. The back- and-forth argument in model theory helps to enlighten the constructions.
