DSpace-CRIS

Now showing 1 - 4 of 4

Some of the metrics are blocked by your
consent settings
Multi-agent deep learning on tensor fields for segmentation of ultrasound images
(2026-03-01)
Sharma, Suman
;
Moodleah, Samart
;
Makhanov, Stanislav S.
Medical image analysis often relies on vector fields (VF), which are fundamental to deterministic models such as Active Contours, Level Set Methods, Phase Portrait Analysis, and artificial agent–based formulations. We experimentally demonstrate that a Deep Learning Neural Network (DLNN) capable of interpreting VF structures can substantially enhance the decision-making capabilities of artificial agents. We introduce a novel hybrid framework that integrates artificial life (AL) agents operating within a VF with a DLNN that guides their behavior. A key innovation of the model is the initialization of AL agents using streamlines derived from the VF orthogonal to the generalized gradient vector flow (GGVF) field. The VF is further transformed into a bi-directional Tensor Field (TF), where the spatial distribution and classification of degenerate points (DPs) serve as critical features. These DPs are leveraged to train AL agents through the DLNN, enabling them to follow meaningful anatomical structures. The framework employs DeepLabV3+ with ResNet50 as the backbone and is trained on 179 benign and 107 malignant breast ultrasound images collected at Thammasat University Hospital (TUH) and annotated by three leading radiologists, in addition to the BUSI and UDIAT datasets. Using 10-fold cross-validation, the proposed method achieves stable and robust performance across three datasets. Mean Dice scores of 94.84±1.63% (TUH), 94.16±1.62% (BUSI), and 93.67±1.51% (UDIAT) are obtained, with corresponding IoU values of 91.19±1.76%, 90.21±1.83% and 89.08±1.70%, demonstrating strong generalization across diverse imaging conditions. Comparative evaluations against state-of-the-art methods confirm the superiority of the proposed model. A video demonstration is available at: https://tinyurl.com/AL-DLNN.
Some of the metrics are blocked by your
consent settings
Edge-Driven Multi-Agent Reinforcement Learning: A Novel Approach to Ultrasound Breast Tumor Segmentation
(2023-12-01)
Karunanayake, Nalan
;
Moodleah, Samart
;
Makhanov, Stanislav S.
A segmentation model of the ultrasound (US) images of breast tumors based on virtual agents trained using reinforcement learning (RL) is proposed. The agents, living in the edge map, are able to avoid false boundaries, connect broken parts, and finally, accurately delineate the contour of the tumor. The agents move similarly to robots navigating in the unknown environment with the goal of maximizing the rewards. The individual agent does not know the goal of the entire population. However, since the robots communicate, the model is able to understand the global information and fit the irregular boundaries of complicated objects. Combining the RL with a neural network makes it possible to automatically learn and select the local features. In particular, the agents handle the edge leaks and artifacts typical for the US images. The proposed model outperforms 13 state-of-the-art algorithms, including selected deep learning models and their modifications.
Some of the metrics are blocked by your
consent settings
Pattern detection for toolpath generation on triangular meshes for 5-axis CNC machining
(2020-01-01)
Dang, Le V.
;
Moodleah, Samart
;
Makhanov, Stanislav S.
Toolpath generation for freeform triangular meshes for 5-axis milling machines using an optimal feeding direction (OFD) field is an important topic in subtractive manufacturing technology. By considering parameters in both the CAD and the CAM stages, there exists a direction(s) at every cutter contact (CC) point, such that the machining efficiency is maximal. The directions could form a special structure which can be detected for toolpath generation on the CAD stage. In this paper, a novel toolpath generation method is proposed to create a toolpath for freeform triangular meshes by detecting regular structures of the OFD in the parametric domain using moment invariants. Three common templates, corresponding to standard toolpath structures, are considered: (1) curl (contour/spiral contour path), (2) star/spiral (radial path), and (3) shear (zigzag path). Mesh parameterization is used to flatten the 3D triangulated surface. A transfinite interpolation (TFI) is utilized to construct the toolpath on the flattened surface. The final toolpath is then obtained by the inverse map. The proposed method has been compared with conventional and standard commercial toolpaths. Numerical results show that the pattern detection algorithm helps to find an appropriate toolpath strategy. The generated toolpaths follow the OFDs closer than the competing methods, and therefore, require shorter machining time.
Some of the metrics are blocked by your
consent settings
Bounded delay performance analysis for distributed multimedia applications under shared virtual environment
(2000-01-01)
Wongwirat, Olarn
;
Makhanov, Stanislav S.
;
Ohara, Shigeyuki
Distributed multimedia applications can be characterized into two types of interaction, nonconversation and conversation, which discrete media and continuous media data are involved respectively. The two applications types have distinct requirement in terms of delay to support interactive services, particularly under shared virtual environment. For conversational interactive type application which continuous media is mainly focused, a limited delay is required for achieving real-time service guarantees and for maintaining temporal relationship, or synchronization. In this paper, we apply the internet delay model to the conversational type application for distributed multimedia data under shared virtual environment on the internet to analyze the performance of the bounded delay value. The simulation results show the delay variation of continuous multimedia data apparently at the clients and a server under dynamic case consideration which represents the delay characteristic of the application on the internet.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results