Nodayama-48 Medeshimashiote, Natori, Miyagi 981-1239, Japan
สำหรับสมาชิกเท่านั้น
I worked on the AR sandbox project. The AR Sandbox uses a computer projector and a motion sensing input device (a kinect 3D Camera) mounted above a box of sand. As a visitor interacts with the sand in the box, the kinect detects the distance to the sand below, and a visualization of an elevation model with contour lines and a color map assigned by elevation is cast from an overhead projector onto the surface of the sand. As visitors move the sand, the kinect perceives changes in the distance to the sand surface, and the projected colors and contour lines change accordingly.
ประวัติการฝึกอบรม
ความสามารถ
ความสามารถทางภาษา
พูด (ดี) อ่าน (ดีมาก) เขียน (ดี)
พูด (พอใช้) อ่าน (ดี) เขียน (ดี)
ไทย 25 คำ/นาที อังกฤษ 32 คำ/นาที
รถยนต์ ,รถจักรยานยนต์ ,
Image processing, Deep learning, Augmented reality, Unity, Python , C#
โครงการ / ผลงาน / เกียรติประวัติ / บุคคลอ้างอิง
โครงงาน
1. Controlling Electronic Devices with Leap Motion(01/2019 - 04/2019) - Electronic devices such as LED can be controlled by Leap motion without touching
2. AR Sandbox(06/2019 - 08/2019) - Kinect and projector produces a topographic map with contour lines that can be created on the sandbox
3. Metal Artifact Recognition using Deep Convolutional Neural Network in Abdomen and Pelvis CT Image(08/2019 - 10/2019) - The study found that the DCNN model can classify images of artifacts that have the highest accuracy as 76.00%. As a result of the dataset that we receive is Thai individual data, which has a relatively small amount of data, resulting in a low accuracy evaluation.
4. Facial Reconstruction from Voice using Deep Learning(10/2019 - 04/2020) - with the goal of creating a human face based on the sound received, with the following and principles: Firstly, an input voice will be passed through the voice embedding network and the generator to draw images, then a discriminator will be used for comparison between output images the reference images will cause the difference of both images called the loss. Send back to the generator to create a new image to reduce the loss until the value is similar to the reference image.
ผลงาน
2018 - The runner-up award at the Accessible Learning Hackathon:
Solving the Right Problems for Students with Disabilities
2019 - 12th Biomedical Engineering International Conference (BMEiCON)