KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 5 of 5

Some of the metrics are blocked by your
consent settings
On-edge 2D-to-3D generative pipeline for seamless instance transformation
(2025-12-01)
Petchhan, Jirayu
;
Doungtap, Surasachai
Despite ongoing challenges with fragmented workflows, latency in device imports, and the main issue of limitations in object reconstruction functionality, relying on imperfect extraction networks remains an impractical solution for scalable object generation. To deal with these constraints, we proposed an end-to-end pipeline that leverages a re-designed self-consistency mechanism—aimed at reducing discrimination, along with the beneficial enhancement from level-set projection and gradient-surface orthogonality. In addition, our approach designs dynamic 3D object creation with minimal manual effort by unifying surface topology and optimizing data loading, enabling a streamlined reconstruction process and more flexible object projection. Our method supports rapid, resource-efficient mesh reconstruction and consistently demonstrates performance improvements across multiple instance benchmarks, covering virtual projection tasks. Improvements in mesh topology reconstruction, as measured by the L1 Chamfer distance (CD) metric, are consistently higher, while the system also achieves significant transmission speedups—up to 56.5×—near-instant importing—along with lowering latency in practical rendering on virtual reality (VR) devices. This result highlights that refining mesh binding improves re-creation fidelity. Our approach to scalability leads to faster user engagement and allows automated deployment without requiring human intervention during importing.
Some of the metrics are blocked by your
consent settings
A Fast and Simple Machine Vision Framework for Approximating the Volume of Axi-Symmetric Objects Using Shadow Ray Casting
(2024-01-01)
Sukprasertchai, Siwakorn
;
Suesut, Taweepol
;
Nunak, Navaphattra
The volume measurement using machine vision system is contactless techniques that play an important role in industries now a day. Basically, three-dimensional reconstruction is required to determine a depth using a special lighting system or multiple cameras. This increases the complexity of the measurement system. A fast and simple machine vision framework called RayVol for estimating the volume of axisymmetric objects in near real-time using a single camera and simple illumination is presented. The RayVol framework employs a shadow casting method to reconstruct the 3D shape of the object by tracing rays from the object’s shadow pixels to the light source location. The result of this technique shows a significant accuracy improvement from the area-projection method. A virtual slice representing the cross-section of an object is reconstructed using a cubic spline approximation from baseline points derived from the boundary pixels of the object image and a shadow casting method. The volume estimation was calculated by restricted integration using the Riemann sum estimation algorithm, and the closed area of the virtual slices was calculated using the shoestring algorithm. Mangoes were used as a case study of the RayVol framework. The volume estimation provides the correlation coefficient of 0.9849 between the developed system and the water replacement method.
Some of the metrics are blocked by your
consent settings
The Influence of Measuring Location on Volume Estimation by Shadow Casting Method
(2024-01-01)
Suesut, Taweepol
;
Sukprasertchai, Siwakorn
Non-contact inspection has dominated various industries for many years, particularly in the field of food and agricultural processes. Volume estimation can be achieved by volumetric imagery or 3D reconstruction. The casting of shadows or shape by shading is an active technique that can be utilized to generate volumetric imagery. By evaluating the length of shadows, the object height data can be achieved, and the transformation of image coordinates to world coordinates is facilitated by means of homography transformation. The width, length, and height of object edges along the entirety of its length are employed to reconstruct virtual slices. The integration of these virtual slices enables the estimation of the object's volume. This research studied the variation of the measuring locations, three different diameters spherical balls 38mm, 50mm, and 80mm are used as test materials. The percentage of errors in volume obtained from all sections by shadow casting were compared to computation volume of sphere formula is approximately -1.6 %, -1.13 %, and - 6.60%, respectively, depending on the size of the object. Furthermore, a visualization-based analysis is presented which has no significant difference. From this studied, it can be concluded that this method is valuable in estimating the volume of axially symmetrical objects.
Some of the metrics are blocked by your
consent settings
A CNN-BASED MULTI-MODEL ENSEMBLE METHOD FOR INDOOR AND OUTDOOR MULTI-VIEW STEREO RECONSTRUCTION
(2022-01-01)
Wattanacheep, Bhattarabhorn
;
Chitsobhuk, Orachat
Camera poses estimation is a critical process that ensures the success of Three-Dimensional (3D) modelling. We present a Convolutional Neural Network (CNN)-based multi-model ensemble method for indoor and outdoor multi-view stereo reconstruction capable of learning across multiple domains, including images from both indoor and outdoor environments. Each domain’s images have distinct properties and shooting view-points, which leads to difficulty in efficient learning such a large difference and requires large amount of computational resources. In order to reduce complexity of the end-to-end single model, the proposed model is divided into multiple learning agents consisting of domain-specific agents and domain relationship agent. The domain-specific agent is trained independently on its own set of unique image characteristics, for example, one for indoor datasets and another for outdoor datasets. The domain relationship agent then ensembles and analyzes the multiple domain features and finalizes the estimation. In terms of average root mean square error, we compare the performance of the combined domain single model with the suggested ensemble CNN model. The experimental results indicate that the proposed model outperforms the others, with rotation and translation prediction errors of 0.112012266.
Some of the metrics are blocked by your
consent settings
Advanced microstructural investigations of AISI 441 early stage oxidation in wet atmosphere
(2017-01-01)
Wongpromrat, Wichitra
;
Parry, Valérie
;
Chandra-Ambhorn, Walairat
;
Chandra-Ambhorn, Somrerk
;
Galerie, Alain
AISI 441 ferritic stainless steel is a good candidate for metallic interconnects in solid oxide fuel cells (SOFCs). The minor elements Ti and Nb are used to stabilize the ferritic matrix and also to reduce creep by a combination of solid solution strengthening and precipitation of intermetallic Laves phase particles along the grain boundaries. However their influence on the oxidation behavior is not well understood. This study focuses on the early stages oxidation (from 4 to 24 h) at 800°C of AISI 441 under 5% H<inf>2</inf>O in O<inf>2</inf>. A relatively smooth micro-crystallized oxide scale and Ti, Nb containing nodules are observed. The internal microstructure of these objects is studied by FIB tomography which allows computing cross sectional views in any direction of interest. FIB study reveals a complex microstructure and a development strongly linked to the presence of niobium and/or titanium in the substrate.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results