Half My Work Ends in a Paper.
The Other Half Ends in Production.

1. Introduction

Since 2023 the work has arrived as several projects in parallel rather than one product, and my research has continued independently alongside it. Both halves are applied computer vision and applied AI, delivered as systems that hold up under real traffic rather than demos. Three domains take most of my time.

Biometric verification. Face recognition, presentation attack and deepfake detection, document and identity checks. Benchmark accuracy is the easy part. What matters is behaviour when someone is actively trying to fool the model [12].

Smart transportation and parking. License plate detection and recognition, vehicle understanding, and the edge deployments that keep a gate moving when the camera is dirty, the light is wrong and the network is not cooperating.

Conversational AI. Assistants wired into the systems a business actually runs on, with authorization enforced at the object level so an agent can never answer with the wrong customer's data.

The research that got me here is on depth from non-conventional cameras: light field depth estimation robust to occlusion and noise [4], [15], cost aggregation for stereo matching [5], [6], and running it all on GPUs, including mobile ones [13], [19]. Other interests include light field synthesis, image enhancement and synthesis [11], and SLAM [14].

2. Background

Williem received a Ph.D. in information and communication engineering from a university in Korea in 2017, working on light field depth estimation, stereo matching and computational photography, and graduated with the Outstanding Researcher Award. Williem then joined the computer science faculty of a private university in West Jakarta, teaching and supervising vision research, before co-founding an AI verification startup as CTO and taking it from nothing to production: the initial technology, the first models, and the engineering team that kept building after. Since 2023, Williem has run an engineering studio that takes on several projects at once, and keeps an independent research practice deliberately separate from it.

3. Service

3.1. Area chair

Area Chair for CVPR 2027.

3.2. Reviewing

Reviewer for the conferences in Table 2, and for the journals IEEE TPAMI, IEEE TIP and IEEE TVCG.

3.3. Organizing

Co-organizer of the Workshop on Computer Vision for Developing Countries (CV4DC), every edition since it started: the 1st with ACCV 2024, the 2nd with ICCV 2025, and the 3rd with ACCV 2026.

4. Recognition

• Outstanding Reviewer, CVPR 2024. Top 2% of 9,872 reviewers.
• Outstanding Reviewer, CVPR 2023. Top 3.5% of 6,625 reviewers.
• Outstanding Reviewer, CVPR 2021. Top 20% of 5,250 reviewers.
• Top 215 Reviewer, ECCV 2020. Top 7.6% of 2,830 reviewers.
• Pioneer Winner, 2020. Global tournament for independent projects.
• Outstanding Research Award, 2017. Dean's choice of the best researcher, graduate school.
• Best Student Paper Award, 2017. Institute of Electronics and Information Engineers, Korea.
• Best Paper Award, 2017. 23rd Japan-Korea Workshop on Frontiers of Computer Vision.
• Best Student Paper Award, 2016. 28th Workshop on Image Processing and Image Understanding.

5. Publications

Complete and current list on Google Scholar. Figure 2 collects teaser figures from the conference papers.

(a) [11] (b) [12] (c) [13] (d) [14] (e) [15] (f) [16] (g) [17] (h) [18] (i) [19]

5.1. Journals

1. Joshua Santoso, Williem, and In Kyu Park. “Holistic 3D body reconstruction from a blurred single image.” IEEE Access, 2022. IF 3.9, h-index 242.
2. Andre Ivan, Williem, and In Kyu Park. “Joint light field spatial and angular super-resolution from a single image.” IEEE Access, 2020. IF 3.9, h-index 242.
3. Jonathan Samuel Lumentut, Williem, and In Kyu Park. “6-DOF motion blur synthesis and performance evaluation of light field deblurring.” Multimedia Tools and Applications, 2019. IF 3.6, h-index 106.
4. Williem, In Kyu Park, and Kyoung Mu Lee. “Robust light field depth estimation using occlusion-noise aware data costs.” IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2018. IF 23.6, h-index 417.
5. Williem and In Kyu Park. “Cost aggregation benchmark for light field depth estimation.” Journal of Visual Communication and Image Representation, 2018. IF 2.6, h-index 96.
6. Williem and In Kyu Park. “Deep self-guided cost aggregation for stereo matching.” Pattern Recognition Letters, 2018. IF 5.1, h-index 181.
7. Williem, Ki Won Shon, and In Kyu Park. “Spatio-angular consistent editing framework for 4D light field images.” Multimedia Tools and Applications, 2016. IF 3.6, h-index 106.
8. Christian Simon, Williem, and In Kyu Park. “Correcting geometric and photometric distortion of document images on a smartphone.” Journal of Electronic Imaging, 2015. IF 1.1, h-index 72.
9. Nicolas Seiller, Williem, Nitin Singhal, and In Kyu Park. “Object-oriented framework for real-time image processing on GPU.” Multimedia Tools and Applications, 2014. IF 3.6, h-index 106.
10. Williem, Yu-Wing Tai, and In Kyu Park. “Accurate and real-time depth video acquisition using Kinect-stereo camera fusion.” Optical Engineering, 2014. IF 1.3, h-index 120.

5.2. Conferences

11. Joshua Santoso, Christian Simon, and Williem. “On manipulating scene text in the wild with diffusion models.” In Proc. IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024.
12. I Putu Agi Karasugi and Williem. “Face mask invariant end-to-end face recognition.” In ECCV Workshop on Real-world Computer Vision from Inputs with Limited Quality, 2020.
13. Andre Ivan, Williem, and In Kyu Park. “Light field depth estimation on off-the-shelf mobile GPU.” In CVPR Workshop on Embedded Computer Vision, 2018.
14. Williem, Andre Ivan, Hochang Seok, Jongwoo Lim, Kuk-Jin Yoon, Ikhwan Cho, and In Kyu Park. “Visual-inertial RGB-D SLAM for mobile augmented reality.” In Proc. Pacific-Rim Conference on Multimedia (PCM), 2017.
15. Williem and In Kyu Park. “Robust light field depth estimation for noisy scene with occlusion.” In Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
16. Williem, Ramesh Raskar, and In Kyu Park. “Depth map estimation and colorization of anaglyph images using local color prior and reverse intensity distribution.” In Proc. IEEE International Conference on Computer Vision (ICCV), 2015.
17. Christian Simon, Williem, Jihwan Choe, Il Dong Yun, and In Kyu Park. “Correcting photometric distortion of document images on a smartphone.” In CVPR Workshop on Mobile Vision, 2014.
18. Williem, Christian Simon, Sungdae Cho, and In Kyu Park. “Fast and robust perspective rectification of document images on smartphone.” In CVPR Workshop on Mobile Vision, 2014.
19. Young-kyu Choi, Williem, and In Kyu Park. “Memory-efficient belief propagation in stereo matching on GPU.” In Proc. APSIPA Annual Summit and Conference, 2012.

5.3. Patents

20. In Kyu Park and Williem. “Method and apparatus for depth image estimation of light fields.” Korean Patent KR101992527B1, 2019. 라이트필드 깊이 영상 추정 방법 및 장치.
21. In Kyu Park and Williem. “Method and apparatus for robust depth image estimation of light fields.” Korean Patent KR101968406B1, 2019. 강인한 라이트필드 깊이 영상 추정 방법 및 장치.
22. In Kyu Park and Williem. “Method and system on deep self-guided cost aggregation for stereo matching.” Korean Patent KR101957089B1, 2019.
23. In Kyu Park, Jun Young Cheong, and Williem. “Method and system for generating anaglyph image reconstruction and depth map.” Korean Patent KR101655036B1, 2016. 애너글리프 영상 복원 및 깊이 지도 생성 방법 및 시스템.

6. Contact

Open to research collaboration, technical advisory, and building things that have to work on the first day they meet real users. Write to [email protected], or find me on Google Scholar and LinkedIn.