Activation Map Networks with Deep Graphical Model for Semantic Segmentation

2019 ◽  
pp. 845-852
Author(s):  
Cheruku Sandesh Kumar ◽  
Ratnadeep Roy ◽  
Sanyog Rawat ◽  
Archek Praveen Kumar
Keyword(s):  
Graphical Model ◽  
Semantic Segmentation ◽  
Activation Map

Optimal Scale of Hierarchical Image Segmentation with Scribbles Guidance for Weakly Supervised Semantic Segmentation

2021 ◽  
pp. 2154026
Author(s):  
Zaid Al-Huda ◽  
Donghai Zhai ◽  
Yan Yang ◽  
Riyadh Nazar Ali Algburi
Keyword(s):  
Image Segmentation ◽  
Graphical Model ◽  
Semantic Segmentation ◽  
Saliency Map ◽  
Training Data ◽  
Deep Convolutional Neural Networks ◽  
High Quality ◽  
Optimal Scale ◽  
Supervised Segmentation ◽  
Weakly Supervised

Deep convolutional neural networks (DCNNs) trained on the pixel-level annotated images have achieved improvements in semantic segmentation. Due to the high cost of labeling training data, their applications may have great limitation. However, weakly supervised segmentation approaches can significantly reduce human labeling efforts. In this paper, we introduce a new framework to generate high-quality initial pixel-level annotations. By using a hierarchical image segmentation algorithm to predict the boundary map, we select the optimal scale of high-quality hierarchies. In the initialization step, scribble annotations and the saliency map are combined to construct a graphic model over the optimal scale segmentation. By solving the minimal cut problem, it can spread information from scribbles to unmarked regions. In the training process, the segmentation network is trained by using the initial pixel-level annotations. To iteratively optimize the segmentation, we use a graphical model to refine segmentation masks and retrain the segmentation network to get more precise pixel-level annotations. The experimental results on Pascal VOC 2012 dataset demonstrate that the proposed framework outperforms most of weakly supervised semantic segmentation methods and achieves the state-of-the-art performance, which is [Formula: see text] mIoU.

scholarly journals SEMANTIC SEGMENTATION OF INDOOR 3D POINT CLOUD WITH SLENET

2019 ◽  
Vol XLII-2/W13 ◽  
pp. 785-791
Author(s):  
Y. Ding ◽  
X. Zheng ◽  
H. Xiong ◽  
Y. Zhang
Keyword(s):  
Point Cloud ◽  
Graphical Model ◽  
Rapid Development ◽  
Semantic Segmentation ◽  
Point Clouds ◽  
3D Point Cloud ◽  
3D Point Clouds ◽  
Cloud Models ◽  
Semantic Transfer ◽  
Class Segmentation

<p><strong>Abstract.</strong> With the rapid development of new indoor sensors and acquisition techniques, the amount of indoor three dimensional (3D) point cloud models was significantly increased. However, these massive “blind” point clouds are difficult to satisfy the demand of many location-based indoor applications and GIS analysis. The robust semantic segmentation of 3D point clouds remains a challenge. In this paper, a segmentation with layout estimation network (SLENet)-based 2D&amp;ndash;3D semantic transfer method is proposed for robust segmentation of image-based indoor 3D point clouds. Firstly, a SLENet is devised to simultaneously achieve the semantic labels and indoor spatial layout estimation from 2D images. A pixel labeling pool is then constructed to incorporate the visual graphical model to realize the efficient 2D&amp;ndash;3D semantic transfer for 3D point clouds, which avoids the time-consuming pixel-wise label transfer and the reprojection error. Finally, a 3D-contextual refinement, which explores the extra-image consistency with 3D constraints is developed to suppress the labeling contradiction caused by multi-superpixel aggregation. The experiments were conducted on an open dataset (NYUDv2 indoor dataset) and a local dataset. In comparison with the state-of-the-art methods in terms of 2D semantic segmentation, SLENet can both learn discriminative enough features for inter-class segmentation while preserving clear boundaries for intra-class segmentation. Based on the excellence of SLENet, the final 3D semantic segmentation tested on the point cloud created from the local image dataset can reach a total accuracy of 89.97%, with the object semantics and indoor structural information both expressed.</p>

Face Completion Using Semantic Segmentation and Geometric Features

2018 ◽  
Vol 11 (6) ◽  
pp. 304
Author(s):  
Javier Pinzon-Arenas ◽  
Robinson Jimenez-Moreno ◽  
Ruben Hernandez-Beleno
Keyword(s):  
Semantic Segmentation ◽  
Geometric Features

DEVELOPMENT OF A METHOD FOR DESIGNING INFORMATION SYSTEMS FOR ENERGY COMPANIES

2020 ◽  
Vol 5 (333) ◽  
pp. 53-57
Author(s):  
T.B. Aldongar ◽  
◽  
F.U. Malikova ◽  
G.B. Issayeva ◽  
B.R. Absatarova ◽  
...  
Keyword(s):  
Information Systems ◽  
Business Processes ◽  
Design Research ◽  
Graphical Model ◽  
Formal Model ◽  
Workflow Management ◽  
Research Process ◽  
Information Models ◽  
The Creation ◽  
Concept Of Information

The creation of information models requires the use of known methods and the development of new methods of formalizing the pre-design research process. The modeling process consists of four stages: data collection on the object of management - pre-project research; creation of a graphical model of business processes taking place in the enterprise; development of a formal model of business processes; business research by optimizing the formal model. To support the creation of workflow management services and systems, the complex offers methodologies, standards and specialized software that make up the developer's tools. This can be ensured only by modern automated methods based on information systems. It is important that the information collected is structured to meet the needs of potential users and stored in a form that allows the use of modern access technologies. Before discussing the effectiveness of FIM, it should be noted that the basic concept of information itself is still not the same. In a pragmatic way, it is a set of messages in the form of an important document for the system. Information can be evaluated not only by volume, but also by various parameters, the most important of which are: timeliness, relevance, value, aging, accuracy, etc. in addition, the information may be clear, probable and accurate. The methods of its reception and processing are different in each case.

A Probabilistic-Graphical-Model Based Approach for Representing Lineages in Uncertain Data

2011 ◽  
Vol 34 (10) ◽  
pp. 1897-1906 ◽  
Author(s):  
Kun YUE ◽  
Wei-Yi LIU ◽  
Yun-Lei ZHU ◽  
Wei ZHANG
Keyword(s):  
Graphical Model ◽  
Uncertain Data ◽  
Probabilistic Graphical Model ◽  
Model Based

Development of environment design support mixed reality system capable of environment estimation using deep learning

Impact ◽  
2020 ◽  
Vol 2020 (2) ◽  
pp. 9-11
Author(s):  
Tomohiro Fukuda
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
Computer Games ◽  
Construction Projects ◽  
Mixed Reality ◽  
Semantic Segmentation ◽  
Environment Design ◽  
Aviation Training ◽  
Architecture And Design ◽  
World Environment

Mixed reality (MR) is rapidly becoming a vital tool, not just in gaming, but also in education, medicine, construction and environmental management. The term refers to systems in which computer-generated content is superimposed over objects in a real-world environment across one or more sensory modalities. Although most of us have heard of the use of MR in computer games, it also has applications in military and aviation training, as well as tourism, healthcare and more. In addition, it has the potential for use in architecture and design, where buildings can be superimposed in existing locations to render 3D generations of plans. However, one major challenge that remains in MR development is the issue of real-time occlusion. This refers to hiding 3D virtual objects behind real articles. Dr Tomohiro Fukuda, who is based at the Division of Sustainable Energy and Environmental Engineering, Graduate School of Engineering at Osaka University in Japan, is an expert in this field. Researchers, led by Dr Tomohiro Fukuda, are tackling the issue of occlusion in MR. They are currently developing a MR system that realises real-time occlusion by harnessing deep learning to achieve an outdoor landscape design simulation using a semantic segmentation technique. This methodology can be used to automatically estimate the visual environment prior to and after construction projects.

Sign in / Sign up

Export Citation Format

Share Document