Research | AIMRL

Research page

Research Methodology: The AIMRL Philosophy

We believe effective machine perception must integrate physical laws with computational intelligence.

At AIMRL, we deconstruct complex automation challenges guided by a systematic pipeline:

Formulation from First Principles: We look beyond existing hardware constraints to solve fundamental physical questions.

Analytical Modeling: We develop deterministic mathematical frameworks for field-based sensing.

Experimental Demonstration: We validate every model through simulation and hardware prototypes to prove technical performance and industrial significance.

The research pages are organized to emphasize transferable principles rather than chronological development. The work presented here represents four decades of exploring the boundary between physical laws and computational intelligence. It is my hope that these findings serve as a foundation

for researchers seeking to transform machine vision from a data-driven pattern recognition task into a robust, measurement-grade science. I encourage the further adaptation of these models to solve the

next generation of challenges in mechatronics and human-centered robotics.

—Kok-Meng Lee

Ⅰ PHYSICS-BASED MACHINE VISION

Thrust 1: Physically-Accurate Machine Perception

1.1 Retroreflective Vision Sensing: Exploiting Directional Gain with Coaxial Optics

1.2 Artificial Color Contrast (ACC): Exploiting Spectral Physics

1.3 Experimental Verification and Industrial Significance:

Case Study A: Engineered Environments

Case Study B: Natural and Industrial Environments

Case Study C: Industrial Processing — ACC–PCA for Poultry Product Segmentation

Thrust 2: Model-Based Machine Vision Systems

2.1 Geometric and Structural Shape Reconstruction