Industrial and Systems Engineering Research Open House Presentation Schedule

Poster ID: 70

1:00 - 2:00

Choi, Joseph

ISE

Grad

Quantitative Texture Characterization of Interstitial Lung Disease using Generative Adversarial Networks

Authors: Choi, Joseph; Chun, Sehyun; Lee, Changhyun; Baek, Stephen*

IPF is a highly lethal and progressive disease that is the class of the ILD. IPF is diagnosed by textures found under HRCT. It is challenging to have consistent and agile diagnosis due to the heterogeneous characteristics of the honeycombing which causes subjective diagnosis. CAD systems were developed to provide a constant diagnosis but still sensitive to the subjective bias in the model. We want to propose an unsupervised method utilizing Generative Adversarial Network to learn the texture manifold of the IPF patients and supply minimal supervision to further guide to draw boundaries of honeycombs and non-honeycombs on the texture manifold.

Three-Minute Video: https://www.youtube.com/watch?v=_5uQyNNeviE

Presentation: choi_ise_vroh_2020.pdf

 

Poster ID: 71

10:00 - 11:00

Fei, Fan

ISE

Grad

Study of Droplet Diffusion in Hydrothermal-Assisted Transient Jet Fusion of Ceramics

Authors: Fei, Fan; He, Li; Kirby, Levi; Xuan Song

Hydrothermal-assisted transient jet fusion (HTJF) is a powder-based additive manufacturing method of ceramics, utilizing a water-mediated hydrothermal mechanism to fuse particles together, eliminating organic binders and thereby contributing to high green-density parts (>90%). Precise control of the liquid diffusion in the powder bed is critical for the fabrication with high density and accuracy. In this research, the dependence of transient solution diffusion on different process parameters were studied. Both numerical modeling and experimental methods were used to quantify the relationships between processing parameters and diffusion profiles. Optimum processing conditions were identified to mitigate the undesired diffusion in the powder bed.

Three-Minute Video: https://youtu.be/jMu82Ew8yDc

Presentation: fei_ise_vroh_2020.pdf

 

Poster ID: 72

9:00 - 10:00

Kim, Heesu

ISE

Grad

Detecting Adequate Use of a Seat Belt for Driver Monitoring Systems

Authors: Kim, Heesu; Chun, Sehyun; Baek, Stephen

We propose a new algorithm to detect adequate use of the seat belt using a convolutional neural network (CNN) combined with a long short-term memory (LSTM) network. The proposed architecture comprises two stages, namely the feature encoding stage and the context generation stage. Processing the two stages, the model classifies the seat belt use into three classes: proper use, improper use, and non use. To validate the proposed algorithm's practical use, we evaluate the model robustness on the lighting condition and driver's appearance and measured the training time and inference time. The result showed that the proposed model is practicable to DMS.

Three-Minute Video: https://www.youtube.com/watch?v=jcn9bnB5PWg

Presentation: kim_ise_vroh_2020.pdf

 

Poster ID: 73

11:00 - 12:00

Murphy, Brandon

ISE

Grad

USABILITY OF COVID-19 DEDICATED PUBLIC HEALTH AGENCY WEBSITES: A HEURISTIC EVALUATION

Authors: Murphy, Brandon; Momenipour, Amir; Rojas-Murillo, Salvador; Pennathur, Priyadarshini; Pennathur, Arunkumar

Amid the COVID-19 pandemic, to communicate accurate, credible, life-saving information to people in a timely manner, government health agencies at all levels in the US, including county, state and federal, have scrambled to create COVID-19 dedicated websites. Because these websites are being used everyday by the general public, by healthcare professionals, and more recently, by decision makers and workers in businesses and educational institutions for making critical decisions, and because it is critical that these websites be usable, we sought to evaluate the usability of these COVID-dedicated government websites in the US.

Three-Minute Video: https://www.youtube.com/watch?v=b_UqNrBCkhQ

Presentation: murphy_ise_vroh_2020.pdf

 

Poster ID: 74

2:00 - 3:00

Whitlow, Harrison

ISE

Undergrad

Detecting Bat Activities at a Wind Farm by Using Infrared Cameras and Deep Neural Networks

Authors: Whitlow, Harrison; Teng, Jian; J. Niemeier, James; Leckband, Jesse; Kruger, Anton; D. Markfort, Corey

Wind energy is a source of renewable energy and is growing remarkably fast. Concerning the rising fatality rate in bats, it is essential to understand a wind turbines effect. As part of broader efforts to monitor bat interactions with wind turbines, this work focuses on the infrared video camera selection criteria, training methods of deep neural network models, and measures for counting bats in post-processing stages of video footage. The results will offer to help establish strategies in moderating wind farm effects on bats.

Three-Minute Video: https://youtu.be/NE2y0N4ev4g

Presentation: whitlow_ise_vroh_2020.pdf