Classification of Valid Test Patterns Using Autoencoder for Adaptive Testing In VLSI
Kunal Nikhade, G Sowmiya, S. Malarvizhi · 2023
As the complexity of the circuit rises during IC manufacture, more test patterns are needed, which extends test time and raises test costs. The number of test patterns needs to be reduced due to the issue of testing time, which is too long for high-input circuits that produce more patterns. In order to pick the valid patterns (patterns that pass the test) and invalid patterns (patterns that fail the test), this study proposes an improved autoencoder classification technique. This classification results in a reduction in the number of test patterns, which decreases test duration. The experimental findings show that the proposed technique classified valid patterns with 86% accuracy.