A General Framework for Context-Specific Image Segmentation Using Reinforcement Learning

2013 ◽  
Vol 32 (5) ◽  
pp. 943-956 ◽  
Author(s):  
Lichao Wang ◽  
K. Lekadir ◽  
S. Lee ◽  
R. Merrifield ◽  
Guang-Zhong Yang
2021 ◽  
pp. 102093
Author(s):  
Antoine Théberge ◽  
Christian Desrosiers ◽  
Maxime Descoteaux ◽  
Pierre-Marc Jodoin

2021 ◽  
Vol 12 (1) ◽  
Author(s):  
Michael C. Krygier ◽  
Tyler LaBonte ◽  
Carianne Martinez ◽  
Chance Norris ◽  
Krish Sharma ◽  
...  

AbstractImage-based simulation, the use of 3D images to calculate physical quantities, relies on image segmentation for geometry creation. However, this process introduces image segmentation uncertainty because different segmentation tools (both manual and machine-learning-based) will each produce a unique and valid segmentation. First, we demonstrate that these variations propagate into the physics simulations, compromising the resulting physics quantities. Second, we propose a general framework for rapidly quantifying segmentation uncertainty. Through the creation and sampling of segmentation uncertainty probability maps, we systematically and objectively create uncertainty distributions of the physics quantities. We show that physics quantity uncertainty distributions can follow a Normal distribution, but, in more complicated physics simulations, the resulting uncertainty distribution can be surprisingly nontrivial. We establish that bounding segmentation uncertainty can fail in these nontrivial situations. While our work does not eliminate segmentation uncertainty, it improves simulation credibility by making visible the previously unrecognized segmentation uncertainty plaguing image-based simulation.


Author(s):  
Yang Gao ◽  
Hao Wang

This chapter concludes three perspectives on multi-agent reinforcement learning (MARL): (1) cooperative MARL, which performs mutual interaction between cooperative agents; (2) equilibrium-based MARL, which focuses on equilibrium solutions among gaming agents; and (3) best-response MARL, which suggests a no-regret policy against other competitive agents. Then the authors present a general framework of MARL, which combines all the three perspectives in order to assist readers in understanding the intricate relationships between different perspectives. Furthermore, a negotiation-based MARL algorithm based on meta-equilibrium is presented, which can interact with cooperative agents, games with gaming agents, and provides the best response to other competitive agents.


Geotechnics ◽  
2021 ◽  
Vol 1 (2) ◽  
pp. 573-587
Author(s):  
Sin Mei Lim ◽  
Linqiao He ◽  
Siang Huat Goh ◽  
Fook Hou Lee

Although there has been a substantial body of research on the chemical stabilization of sewage sludge, most of these results are project-specific and relate mainly to the use of new binders and sewage sludge from specific sources. In this sense, much of the work to date is context-specific. At present, there is still no general framework for estimating the strength of the chemically treated sludge. This paper proposes one such general framework, based on data from some recent studies. An in-depth re-interpretation of the data is first conducted, leading to the observation that sludge, which has coarse, hard particulate inclusions, such as sand, premixed into it, gives significantly higher strength. This was attributed to the hard coarse particles that lower the void ratio of treated soil, are much less susceptible to volume collapse under pressure, and contribute to the strength through frictional contacts and interlocking. This motivates the postulation of a general framework, based on the premise that coarse, hard particulate inclusions in the sludge which do not react with the binders can nonetheless contribute to the strength of the treated soil. The overall void ratio, defined as the volume of voids in the cementitious matrix normalised by the overall volume, is proposed as a parameter for quantifying the combined effect of the coarse particulate inclusions and the cementitious matrix. The binder-sludge ratio is another parameter which quantifies the strength of the cementitious matrix, excluding the hard particulate inclusions. Back-analysis of the data suggests that the significance of the binder-sludge ratio may diminish as the content of hard particulate inclusions increases.


Author(s):  
Alpa Singh Rajput ◽  
S. S. Thakur ◽  
Om Prakash Dubey

Purpose: In the present paper the concept of soft almost β-continuous mappings and soft almost β-open mappings in soft topological spaces have been introduced and studied. Methodology: This notion is weaker than both soft almost pre-continuous mappings, soft almost semi-continuous mapping. The diagrams of implication among these soft classes of soft mappings have been established. Main Findings: We extend the concept of almost β-continuous mappings and almost β-open mappings in soft topology. Implications: Mapping is an important and major area of topology and it can give many relationships between other scientific areas and mathematical models. This notion captures the idea of hanging-togetherness of image elements in an object by assigning strength of connectedness to every possible path between every possible pair of image elements. It is an important tool for the designing of algorithms for image segmentation. The novelty of Study: Hope that the concepts and results established in this paper will help the researcher to enhance and promote the further study on soft topology to carry out a general framework for the development of information systems.


Sign in / Sign up

Export Citation Format

Share Document