The Patent Landscape: 40+ Filings and a Post-2020 Surge in AI-Driven PCB AOI

The patent record for AI-driven PCB automated optical inspection encompasses more than 40 active, pending, or recently granted patents and technical disclosures filed between 2004 and 2026, with the highest concentration of filings occurring after 2020. This acceleration maps directly onto the broader availability of large pre-trained vision models and the maturation of transfer learning frameworks that make domain adaptation tractable without building inspection models from scratch.

The dominant technical approaches across the dataset are: (1) transfer learning and fine-tuning of pre-trained vision models — CNNs, Vision Transformers, and CLIP-based architectures — onto PCB-specific defect datasets; (2) hybrid AOI pipelines where optical image processing provides initial defect flags that a trained adaptive algorithm refines or overrides; (3) cross-modal prompt learning combining visual and text encoders; and (4) continual learning for incremental defect type expansion. A secondary but significant cluster covers YOLO-family model adaptations for real-time PCB defect detection. Dominant assignees span both industrial OEMs and academic research groups: Siemens AG, Tsinghua University, Central South University, Robert Bosch GmbH, Suzhou Kangdai Intelligent Technology, and United Automotive Electronics.

Hybrid AOI Pipelines: Foundation Models as Second-Stage Classifiers

The most architecturally significant innovation in recent PCB AOI patents is the replacement — or augmentation — of classical rule-based optical image processing with a trained adaptive algorithm acting as a second-stage decision-maker. Rather than relying solely on threshold-based comparisons, hybrid pipelines use a pre-trained machine learning model to re-evaluate candidate defects flagged by the optical subsystem, directly addressing the chronic problem of false positives that previously required operator-intensive manual review and test-program maintenance after every process or lighting change.

Siemens AG is the most prolific patent filer in this architectural category, with at least three major PCB/PCBA AOI patents filed between 2023 and 2024 — all centered on the trained adaptive algorithm layered on top of classical optical inspection. The Method for Optical Quality Control During Circuit Board Manufacturing (Siemens, 2023) defines a workflow where a first defect indicator is produced by conventional optical image processing, and only when this indicator flags a defect is a second defect indicator produced by a trained adaptive ML algorithm. The output of the ML model is then substituted for the optical result. A complementary Siemens patent, Inspection of Printed Circuit Board Assemblies (Siemens, 2024), further specifies that AOI numerical measurement results — extracted as feature vectors with dimensionality reduction — can be fed into a classifier trained on image-label pairs for detecting false calls in SMT manufacturing.

“A shared visual foundation model generates a universal feature representation, while multiple task-specific detection sub-models are fine-tuned per production line domain — the most explicit treatment of foundation model fine-tuning in the dataset.”

Robert Bosch GmbH’s approach, disclosed in Method and Apparatus for Automated Optical Inspection (Robert Bosch, 2025), introduces a domain-specific decomposition: a shared visual foundation model — described as a hierarchical Vision Transformer backbone using architectures such as Swin, CvT, CSwin, SAM, or DINOv2 — generates a universal feature representation, while multiple task-specific detection sub-models are fine-tuned per production line domain. The patent explicitly distinguishes between a “general feature extraction sub-model” that is pre-trained and frozen, and task detection sub-models that are domain-adapted. Each domain-specific sub-model is trained on images and labels collected from its specific production environment, making fine-tuning both targeted and computationally efficient.

Siemens’ X-ray inspection patent (System and Method for Automatic X-Ray Inspection, Siemens, 2023) extends the same principle to 3D convolutional neural networks trained on ground-truth labeled X-ray image stacks — including cross-sectional slice sequences — to classify solder-joint defects such as pin shorts. According to WIPO, patent filings in AI-assisted manufacturing inspection have grown substantially since 2019, consistent with the post-2020 surge observed in this dataset. The architectural pattern of a frozen backbone with fine-tuned task heads aligns with established transfer learning principles documented by IEEE in the context of industrial computer vision.

Cross-Modal Prompt Learning and Vision-Language Fine-Tuning for PCB Anomaly Detection

A newer and technically advanced category of fine-tuning approaches draws on vision-language foundation models — architectures pre-trained on paired image-text data — and adapts them to PCB anomaly detection through prompt engineering and cross-modal alignment. These methods are particularly valuable because they can operate effectively under data-sparse conditions, a key bottleneck in PCB AOI given the rarity of real defect samples in production environments.

Central South University filed two closely related patents in 2025 covering this paradigm. The PCB Defect Detection Method and System Based on Cross-Modal Prompt Learning and Visual Guidance (Central South University, 2025) constructs a model with a pre-trained visual encoder and a pre-trained text encoder as the feature extraction backbone. Normal and anomalous PCB states are encoded as text prompts, and the model learns a joint cross-modal feature space. During fine-tuning, a dual-constraint alignment strategy combining contrastive loss and triplet loss adjusts the mapping between visual and text features to establish a clear normal/anomaly boundary. The system also introduces a dual-channel repository mechanism with hard-negative sample mining: synthetic anomaly samples generated from defect-free PCB images using structure-and-physics-prior augmentation are screened to select the most training-informative examples.

A second Central South University patent refines the inference pipeline: the visual encoder’s spatial feature vectors are compared against stored text feature embeddings at every spatial position to generate similarity maps, attention maps, or heat maps, which are then upsampled and thresholded to produce binary defect localization masks. This approach is explicitly claimed to improve accuracy, robustness, and generalization under data-sparse conditions.

Zhejiang University extends cross-modal reasoning to a PCB-specific anomaly detection network, PCNet, in PCB Multi-Level, Multi-Scale Anomaly Detection Network PCNet (Zhejiang University, 2025). The method inputs PCB images into a frozen CLIP model’s visual encoder to obtain initial visual embedding feature maps, then passes them through a series of prompt convolution blocks that effectively fine-tune the response of the frozen foundation model without full retraining. Synthetic anomalous images with labels are generated from normal samples and used as training data, producing a system capable of effective anomaly detection with only a small number of normal training examples. Research published by Nature on few-shot learning in industrial inspection contexts corroborates the practical value of prompt-based fine-tuning when labeled defect data is scarce.

Transfer Learning, Data Augmentation, and Continual Learning for Deployed AOI Models

A large cluster of patents addresses the practical challenge of limited labeled defect data by combining pre-trained model fine-tuning with data augmentation and active or continual learning strategies. These approaches tackle two distinct but related problems: bootstrapping accurate models when real defect samples are rare, and keeping deployed models current as PCB designs and manufacturing processes evolve.

Jiangxi Huashi Optoelectronics’ AOI Defect Detection Method and System Based on Image Processing (2026) explicitly integrates transfer learning and meta-learning into the fine-tuning pipeline. A conditional GAN generates defect-free reference images as comparison baselines, while spatial attention mechanisms and multi-scale feature pyramid networks are combined with a pre-trained defect classification model that is fine-tuned using transfer learning and meta-learning frameworks. The patent claims this combination significantly reduces false-positive and missed-detection rates while improving adaptability and generalization across complex texture backgrounds.

Tsinghua University has been the most productive Chinese academic assignee, with at least four patents from 2020 to 2023. The Real-Time Automatic Detection Device for 2D PCB Defects Based on Deep Learning (Tsinghua University, 2020) establishes a cloud-connected CNN architecture encompassing defect training, detection, classification, and filtering modules. CNN models are trained on a labeled PCB defect library containing multi-scale images organized as standard/defect image pairs; the cloud deployment allows centralized model updates without local re-programming. This was extended in 2022 to support three detection modes: supervised comparison-based CNN, unsupervised (no-reference) CNN, and a hybrid mode — enabling the same model to handle both known and novel defect types.

United Automotive Electronics filed two closely related patents (2021; 2024) describing a re-inspection model fine-tuned by deep learning algorithms on AOI output files. The model selects the correct re-inspection sub-model based on the imaging algorithm associated with each PCB image type, outputs confidence scores and heat maps, and uses an active learning strategy to optimize the model on partially labeled samples. The use of Maximum Mean Discrepancy (MMD) and KL divergence as distribution metrics enables domain adaptation between labeled and unlabeled samples — a form of semi-supervised fine-tuning. The University of Science and Technology of China (USTC) addresses few-shot fine-tuning directly in its 2023 patent: an instance-level embedding backbone trained with position-weight and sample-weight assignment modules learns defect-class-aware prototypes from only a small number of labeled examples, enabling recognition of rare defect categories without large-scale retraining. These approaches align with few-shot learning frameworks surveyed by OECD in its analysis of AI adoption in manufacturing quality control.

Key Players: Industrial OEMs, Academic Pioneers, and the Bifurcation of Innovation

The patent data reveals a clear bifurcation between industrial OEM players who focus on production-ready hybrid pipelines and research institutions that advance the algorithmic frontiers of fine-tuning. Understanding where each player sits in this landscape is essential for R&D teams benchmarking their own AOI development roadmaps.

IBM introduced a specialized hardware-assisted fine-tuning approach in Visual Inspection Training Board for Artificial Intelligence Deep Learning (IBM, 2024): physical PCBs with intentionally embedded defects are used as dedicated training artifacts for fine-tuning AI algorithms in the AOI system. This standardized, reproducible approach bootstraps AI model training without relying on rare production defect samples — a practical solution to the data scarcity problem that complements purely algorithmic approaches such as GAN-based augmentation. IBM’s approach can be explored further in the context of PatSnap’s IP intelligence solutions for hardware and manufacturing domains.

Suzhou Kangdai Intelligent Technology filed a PCB defect detection patent (2023) implementing a cascaded multi-sub-model architecture where each sub-model is pre-trained on a specific defect type, and mismatched predictions trigger human annotation followed by retraining — implementing a supervised fine-tuning feedback loop directly in the production environment. YOLO-family model fine-tuning is the most frequently cited detection architecture across smaller Chinese university assignees including Shanghai University, HDUET, Huaiyin Institute of Technology, Northwest Polytechnical University, and Guangzhou Railway Polytechnic, all reporting improvements in mAP, recall, and false-negative rate through structural modifications combined with fine-tuning on PCB-specific datasets. For broader context on AI adoption in manufacturing quality systems, PatSnap Insights regularly covers patent intelligence across industrial AI domains.