AOI-SSL: Self-Supervised Framework for Efficient Segmentation of Wire-bonded Semiconductors In Optical Inspection

TL;DR

AOI-SSL introduces a self-supervised, efficient segmentation framework for wire-bonded semiconductor inspection, reducing labeled data needs and enabling rapid adaptation to new devices.

Contribution

The paper presents a novel self-supervised training approach combining vision transformer pre-training and in-context retrieval for semiconductor segmentation.

Findings

Masked Autoencoders outperform other self-supervised methods in small-data settings.

Retrieval-based segmentation matches complex attention methods with minimal training.

Self-supervised pre-training enhances segmentation quality over scratch or ImageNet pre-training.

Abstract

Segmentation models in automated optical inspection of wire-bonded semiconductors are typically device-specific and must be re-trained when new devices or distribution shifts appear. We introduce AOI-SSL, a training-efficient framework for semantic segmentation of wire-bonded semiconductors by combining small-domain self-supervised pre-training of vision transformers with in-context inference that minimizes the need of labeled examples. We pre-train SOTA self-supervised algorithms in a small industrial inspection dataset and find that Masked Autoencoders are the most effective in this small-data setting, improving downstream segmentation while reducing the labeled fine-tuning effort. We further introduce in-context, patch-level retrieval methods that predict masks directly from dense encoder embeddings with negligible additional training. We show that, in this setting, simple similarity-based retrieval performs on par with more complex attention-based aggregation used currently in the literature. Furthermore, our experiments demonstrate that self-supervised pre-training significantly improves segmentation quality compared to training from scratch and to ImageNet pre-trained backbones under a fixed fine-tuning computational budget. Finally, the results reveal that retrieval based segmentation outperforms fine-tuning when targeting single device images, allowing for near-instant adaptation to difficult samples.