AI R&D Engineer at ASUS

Chi-Mao (Jonathan) Fan

Computer vision.

I develop AI algorithms and integrate them into practical software, from model training and evaluation to backend services and frontend interfaces.

View Competitions GitHub LinkedIn

Off the clock

Piano & Guitar
Card Magic
Hiking & Climbing
AI & Computer Vision
Reading & Writing
Skateboarding
Stargazing
Photography
Anything & Everything

Action-figure style illustration of Jonathan with coffee cups, playing cards and a guitar

Scroll

About

I am an AI R&D Engineer at ASUS, specializing in computer vision and turning machine learning models into practical software products. My work spans algorithm development, model evaluation, backend services, and frontend interfaces.

My experience covers AISVision R6, AISVision 365, and AISDetector 2.0, including image classification, segmentation, object detection, anomaly detection, tracking, and signal analysis. A key focus is oriented and directional object detection, including 360° orientation estimation and model evaluation.

I enjoy connecting research with engineering, from understanding a problem and validating a model to integrating it into a usable application.

I hold a master’s degree from National Chung Hsing University, where my research focused on deep learning for image restoration. My work has appeared at ICIP, EUSIPCO, and ISCAS, and my thesis received the 2022 IEEE Taipei Section Best Dissertation Award. I am also the lead inventor on four granted utility model patents in Taiwan.

I take part in AI competitions, read and share research papers, and build internal tools that improve the team’s debugging workflow. I am also exploring LLMs, retrieval-augmented generation, and development automation.

Publications: 7
Patents: 4
Competitions: 8+
GitHub stars: 597

Focus areas

Computer Vision
Deep Learning
Oriented Object Detection
Image Restoration
Anomaly Detection
Object Tracking
Signal Analysis
AI Product Integration

Core stack

Python
PyTorch
FastAPI
PostgreSQL
Vue 3
TypeScript

Résumé

Experience

ASUSTeK COMPUTER INC.

AI R&D Engineer

Digital Image Technology (DIT) Product Department · Taipei, Taiwan · Since September 2022

2022 – Present

I develop AI algorithms and integrate them into practical software, from model training and evaluation to backend services and frontend interfaces. The years below are the periods I mainly worked on each product.

Products
1. AISDetector 2.0 2026–Present
  
  Signal analysis and preprocessing
  - Backend services and model architectures for signal analysis.
  - CI/CD and benchmarking for model releases, plus a preprocessing recommendation prototype.
2. AISVision 365 2025–2026
  
  AI model and product integration
  - Frontend and later backend for model training, inference, reporting, and verification.
  - Oriented and directional object detection, dataset conversion tools, and packaging.
3. AISVision R6 2022–2025
  
  Computer vision and algorithm development
  - Object detection with 360° orientation estimation: angle-aware losses, rotated NMS, and metrics.
  - Classification, segmentation, anomaly detection, and tracking algorithms.
Other contributions
- Built internal log-window tools that speed up R&D debugging for AISVision R6 and AISDetector 2.0.
- Share research papers with the team and take part in AI competitions.
- Lead inventor on four granted utility model patents in Taiwan.
- Python
- PyTorch
- FastAPI
- PostgreSQL
- Vue 3
- TypeScript
- CI/CD
ITRI

Android System Software Development Intern

2021 – 2022

In my second year of the master's program, I received an interview invitation from the Industrial Technology Research Institute's Information and Communications Labs, where I was responsible for supporting the Android system development team. I learned and practiced the technique of flashing Android image files onto development boards, enhancing my understanding of the Android operating system architecture and the application installation process.
- C
- Android Development
- Embedded Systems

Publications

ICIP 2023

Compound Multi-branch Feature Fusion for Real Image Restoration

IEEE International Conference on Image Processing (ICIP 2023)

Abstract

Image restoration is a challenging and ill-posed problem which also has been a long-standing issue. However, most of learning based restoration methods are proposed to target one degradation type which means they are lack of generalization. In this paper, we proposed a multi-branch restoration model inspired from the Human Visual System (i.e., Retinal Ganglion Cells) which can achieve multiple restoration tasks in a general framework. The experiments show that the proposed multi-branch architecture, called CMFNet, has competitive performance results on four datasets, including image deblurring, dehazing and deraindrop which are very common applications for autonomous cars.

Paper arXiv Code 100 Video
Master’s Thesis Best Dissertation Award

Image Restoration using Improved Hierarchical Encoder-Decoder Networks with Selective Residual Blocks

Master’s thesis, NCHU · 2022 IEEE Taipei Section Best Dissertation Award

Abstract

In this paper, we based on light hierarchical network architecture: U-Net, and improve from Residual Dense Block (RDB) which is good at image restoration tasks but memory-consuming to an efficient block called Selective Residual Block (SRB). We also improve the hierarchical network structure U-Net by adding the gatepost feature paths which enrich more spatial feature information comparing with the traditional U-Net and have the synergy with SRB. Besides this, we also proposed a loss function which is based on two important metrics in image restoration: peak signal-to-noise (PSNR) and structural similarity index to optimize our model. Finally, proposed network could handle the 9 different restoration tasks such as denoising, deblurring, deraining, dehazing and low-light image enhancement. Furthermore, the performances are good in terms of quantitative metrics and visual quality. This paper was selected for the IEEE 2022 Taipei Section Best Master's and Ph.D. Thesis Award 🎉

Paper Code 12 Certificate
ICIP 2022

Half Wavelet Attention on M-Net+ for Low-Light Image Enhancement

IEEE International Conference on Image Processing (ICIP 2022)

Abstract

Low-Light Image Enhancement is a computer vision task which intensifies the dark images to appropriate brightness. It can also be seen as an illposed problem in image restoration domain. With the success of deep neural networks, the convolutional neural networks surpass the traditional algorithm-based methods and become the mainstream in the computer vision area. To advance the performance of enhancement algorithms, we propose an image enhancement network (HWMNet) based on an improved hierarchical model: M-Net+. Specifically, we use a half wavelet attention block on M-Net+ to enrich the features from wavelet domain. Furthermore, our HWMNet has competitive performance results on two image enhancement datasets in terms of quantitative metrics and visual quality.

Paper Code 108 Video
ICCE-TW 2022

Improved Hierarchical M-Net+ for Blind Image Denoising

IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW 2022)

Abstract

Image denoising is a long standing ill-posed prob-lem. Recently, the convolution neural networks (CNNs) gradually stand in the spotlight and almost dominated the computer vision field and had achieved impressive results in different levels of vision tasks. One of famous hierarchical CNN-backbones is the U-Net which shows awesome performance in both denoising and other areas of computer vision. However, the hierarchical architecture usually suffers from the loss of spatial information due to the repeated sampling. It seriously affects the denoising performance especially the element-wise task like denoising. In this paper, we proposed an improved hierarchical backbone: M-Net+ for image denoising to ameliorate the loss of spatial details. Furthermore, we test on two synthetic Gaussian noise datasets to demonstrate the competitive result of our model.

Paper
EUSIPCO 2022

Selective Residual M-Net for Real Image Denoising

IEEE European Signal Processing Conference (EUSIPCO 2022)

Abstract

Image restoration is a low-level vision task which is to restore degraded images to noise-free images. With the success of deep neural networks, the convolutional neural networks surpass the traditional restoration methods and become the main-stream in the computer vision area. To advance the performance of denoising algorithms, we propose a blind real image denoising network (SRMNet) by employing a hierarchical architecture improved from U-Net. Specifically, we use a selective kernel with residual block on the hierarchical structure called M-Net to enrich the multi-scale semantic information. Furthermore, our SRMNet has competitive performance results on two synthetic and two real-world noisy datasets in terms of quantitative metrics and visual quality. The source code and pretrained model are available at https://github.com/FanChiMao/SRMNet.

Paper Code 31 Video
ISCAS 2022

SUNet: Swin Transformer UNet for Image Denoising

IEEE International Symposium on Circuits and Systems (ISCAS 2022)

Abstract

Image restoration is a challenging ill-posed problem which also has been a long-standing issue. In the past few years, the convolution neural networks (CNNs) almost dominated the computer vision and had achieved considerable success in different levels of vision tasks including image restoration. However, recently the Swin Transformer-based model also shows impressive performance, even surpasses the CNN-based methods to become the state-of-the-art on high-level vision tasks. In this paper, we proposed a restoration model called SUNet which uses the Swin Transformer layer as our basic block and then is applied to UNet architecture for image denoising.

Paper arXiv Code 265 Video
ICCE-TW 2021

WBTP-VTON: Whole Body and Texture Preservation based Virtual Try-On Network

IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW 2021)

Abstract

Image-based virtual clothes try-on systems are becoming more and more popular. However, many challenges are waiting to be solved. Therefore, we propose a new fully learnable method, called the whole body and texture preservation based virtual try-on network (WBTP-VTON) to guide the virtual attempt to deal with all practical challenges in this area. First, the WBTP-VTON template conversion is used to transform the target clothing and pants (or skirts) according to the body shape of the target person using a method called Geometric Matching Module (GMM). The second part is to synthesize the final image and make the generated results more realistic. Finally, we use try-on modules and synthetic masks to combine the deformed clothes and the final image to ensure image smoothness. After experimenting on a large data set, it is proved that our WBTP-VTON method has advanced virtual try-on performance.

Paper Code 5

Drag to compare

Image restoration results from my thesis

All comparisons

Input Restored

Kodak24 Drag the slider or use the arrow keys.

Patents

Lead inventor on four granted utility model patents in Taiwan. The applicant and patent holder is ASUSTeK COMPUTER INC.

Taiwan utility model

Electronic device

Lead inventor

Publication no.

TWM681536U

Application no.

TW114210536

Filed

Oct 2, 2025

Published

Apr 1, 2026

Applicant

ASUSTeK COMPUTER INC.
Taiwan utility model

Electronic device

Lead inventor

Publication no.

TWM678561U

Application no.

TW114210089

Filed

Sep 19, 2025

Published

Dec 21, 2025

Applicant

ASUSTeK COMPUTER INC.
Taiwan utility model

Digital audio system

Lead inventor

Publication no.

TWM653962U

Application no.

TW112211484

Filed

Oct 24, 2023

Published

Apr 11, 2024

Applicant

ASUSTeK COMPUTER INC.
Taiwan utility model

Electronic device for audio transcription

Lead inventor

Publication no.

TWM652294U

Application no.

TW112211968

Filed

Nov 3, 2023

Published

Mar 1, 2024

Applicant

ASUSTeK COMPUTER INC.

Competitions

2026

RepoRadar: AI Issue Intelligence

ASUS internal AI Agent competition. An Electron desktop app that brings GitHub and GitLab issues, discussions, and related PRs/MRs into one workspace, with AI summaries, RAG chat, analytics, scheduled sync, and Markdown/HTML/PDF reports. TypeScript frontend, Python FastAPI backend.

GitHub 1 Demo

2024

Low-Frame-Rate Vehicle Tracking

Team result: 2nd of 286 teams

AI CUP 2024. Contributed to the ASUS team's solution for tracking vehicles in 1 fps video, combining YOLO detector ensembles, OSNet re-identification features, and motion-based association.

GitHub 29 Certificate

2024

RAG for Financial Q&A

AI CUP 2024, as part of TEAM_6029. Built a reproducible workflow that retrieves the relevant evidence from financial documents, with baseline and submission pipelines and precision@1 evaluation.

GitHub 1

2023

Shuttlecock Tracking

Team result: top 25% of 439 teams

Used AI methods to locate the shuttlecock and analyze its trajectory in badminton match videos.

GitHub 9 Certificate

2022

CyberAudio: Music Transcription

Team result: 3rd of 60 teams

Intel DevCUP 2022. Combined source separation, music transcription, and OpenVINO to separate the instruments in a song and produce MIDI and sheet music.

GitHub 6 Certificate

2022

Orchid Species Classification

Team result: 18th of 275 teams

AI CUP 2022. Combined 7 classification models with ensemble methods to predict orchid species, and received a Project Office Award Certificate.

GitHub 1

2021

Data Regression

Top 25% of 118 teams

Predicted industrial data with a multilayer perceptron, then retrained the model on site in real time with newly provided training data.

GitHub 4

2021

Rice Plant Detection

9th of 523 teams · Honorable Mention

AI CUP 2021, organized by the Ministry of Education and Academia Sinica. Located rice plants in UAV RGB imagery with a YOLOv4 detector (university division).

GitHub 5

Education

2020–2022

Master of Science (M.S.), Electrical Engineering

National Chung Hsing University (NCHU)

Research focus: computer vision, deep learning, and image restoration, including denoising, deblurring, raindrop removal, dehazing, and low-light enhancement. Papers at ICIP, EUSIPCO, and ISCAS. GPA 4.13/4.15. My thesis received the 2022 IEEE Taipei Section Best Dissertation Award.

Sep 2016–Jun 2020

Bachelor of Science (B.S.), Electrical Engineering

Yuan Ze University (YZU)

Coursework in electronics, circuits, and electromagnetics, plus data structures and C programming. My capstone was a classroom interaction Android app built with socket networking and SQLite. That project is where I found I enjoy software more than hardware.