INSPECTION & TEST

Inspection & test equipment

Wafer Inspection & Test

Metrology, defect inspection and electrical test across the whole line.

EQUIPMENT SCOPE

Equipment scope

Focused on leading Japanese brands, covering electrical test and optical or laser inspection equipment.

SERVICES

AI INTEGRATION

The third-party frontier architectures BETTER benchmarks and deploys.

G2LGAN + CNN (2025)

Two-stage GAN augmentation with MobileNetV2 handles class-imbalanced wafer map defect classification.

Benchmark metric: accuracy 98.39%, F1 93.01%

Benchmark architecture

CNN-ESN (2026)

ResNet34 with an echo state network improves recognition robustness on noisy wafer maps.

Benchmark metric: 94.74% clean, 87.30% noisy

Benchmark architecture

SEMVision H20 (2025)

Deep learning image classification, trained continuously in the fab, speeds up e-beam defect review.

Benchmark metric: 3× review speed, deployed at 2nm/GAA customers

Benchmark architecture

GPU real-time adaptive test

GPU-accelerated machine learning adapts the test flow in real time and shortens test time.

Benchmark metric: Blackwell production, test time down 25%

Benchmark architecture

KGD and outlier detection (GPR, conformal prediction)

Spatial modeling and conformal prediction sharpen known good die screening and outlier detection.

Benchmark metric: outperforms DPAT, about 90% Vmin coverage guarantee

Benchmark architecture

InF-ATPG (2025)

FFR partitioning combined with QGNN and DQN optimizes test pattern generation.

Benchmark metric: backtracks down 55.06%, UFP 0.50%

Benchmark architecture

USE CASES

GET IN TOUCH

One point of contact from equipment selection through installation and AI integration.