Shuning Jiang

Ph.D student
Computer Science and Engineering
The Ohio State University
Office: 680 Dreese Lab
Email: jiang.2126@osu.edu
LinkedIn: jiangsn

I am a Ph.D student in Computer Science & Engineering Department at The Ohio State University. Currently, I am working in Interactive Visual Computing Lab (IVCL), with Prof. Jian Chen. I am interested in human-AI collaboration in interdisciplinary science and skilled in visualization, data analysis, deep learning, and web development.


Projects Portfolio


Pathologiests' gaze analysis
(JavaScript, OpenSeadragon, Python, PyTorch)
◦ Collected participants’ eye movements and mouse operations when making diagnoses on pathology images. Designed an algorithm to generate pseudo ground truth from gaze data for neural network training.


Brain imaging visualization and annotation analysis [GitHub]
(C++, Qt 5, OpenGL, OpenVR)
◦ Studied the annotators’ behaviors in both desktop and virtual reality environments. Used artificial neural networks to boost annotators’ speed and fasten the neuron reconstruction process. In cooperation with Allen Institute and Southeast University.


Image Complexity Quantification [Link]
(jsPsych, JavaScript, HTML/CSS, PHP, MySQL)
◦ Evaluated convolutional neural networks’ performance on graphical perception tasks and compared it to human fairly. Quantified the effect of data sampling on CNNs’ accuracy.


Graphical perception with CNNs
(Python, TensorFlow, Keras)
◦ Evaluated convolutional neural networks’ performance on graphical perception tasks and compared it to human fairly. Quantified the effect of data sampling on CNNs’ accuracy.


Similarity-based image retrieval in pathology
(Python, TensorFlow, Keras, JavaScript, D3.js)
◦ Benchmarked multiple image feature extraction algorithms and retrieved similar images in the database.


Online vape shop analysis
(Python, Web Scraper)
◦ Automatically scraped and mined vape product information from online vape shops.


Automatic vehicles detection and recognition system
(Python, PyTorch, OpenCV, MySQL)
◦ Used convolutional neural networks to automatically detecting and re-identifying vehicles based on license plate, vehicle model, and color.


Publications

Are CNNs More Effective than Humans for Ratio Estimate Tasks? [Download poster] [Online Study]
Shunning Jiang, Wei-Lun Chao, Jian Chen, Daniel Haehn, Meng Ling, Ce Shang, and Hanspeter Pfister.
Vision Sciences Society (VSS), poster presentation, 2022
Summary ◦ Be the first to compare humans and AI on graphical reading tasks. We demonstrated that CNNs behaved differently from humans but can still achieve super-human accuracy with the sample space been covered by the training data.


Excise Taxes and Pricing Activities of E-liquid Products Sold in Online Vape Shops
Shaoying Ma, Shuning Jiang, Meng Ling, Bo Lu, Jian Chen, and Ce Shang.
Tobacco Control, 2022
Summary ◦ We automatically scraped and cleaned online e-liquid pricing and excise taxes activities. We found the e-liquid products sold online were priced low and stricter enforcement of e-cigarette excise tax is needed.

Price Promotions of E-liquid Products Sold in Online Stores
Shaoying Ma, Shuning Jiang, Meng Ling, Jian Chen, and Ce Shang.
International Journal of Environmental Research and Public Health, 2022
Summary ◦ We automatically scraped and cleaned online e-liquid pricing and excise taxes activities. We found 92.36% of the eliquid products were on sales and the effect of excise tax may be offset by price promotion activities.

Employment

The Ohio State University
Research Assistant working with Prof. Jian Chen. and Prof. Ce Shang.

Teaching Assistant for
- CSE 6431 - Advanced Op System
- CSE 5546 - Virtual Reality
- CSE 5544 - Intr Data Vis
- CSE 5052 - AI for Non Majrs
- CSE 3521 - Intr Artif Intell




Online Experiments

Bar chart ratios as far as the eyes can see [Link]
Are different bar chart styles created equal? Find out by taking quick estimates of ratios from bar charts. This test tasks about 5 minutes.