A Novel Method to Inspect 3D Ball Joint Socket Products Using 2D Convolutional Neural Network with Spatial and Channel Attention

Abstract

Product defect inspections are extremely important for industrial manufacturing processes. It is necessary to develop a special inspection system for each industrial product due to their complexity and diversity. Even though high-precision 3D cameras are usually used to acquire data to inspect 3D objects, it is hard to use them in real-time defect inspection systems due to their high price and long processing time. To address these problems, we propose a product inspection system that uses five 2D cameras to capture all inspection parts of the product and a deep learning-based 2D convolutional neural network (CNN) with spatial and channel attention (SCA) mechanisms to efficiently inspect 3D ball joint socket products. Channel attention (CA) in our model detects the most relevant feature maps while spatial attention (SA) finds the most important regions in the extracted feature map of the target. To build the final SCA feature vector, we concatenated the learned feature vectors of CA and SA because they complement each other. Thus, our proposed CNN with SCA provides high inspection accuracy as well as it having the potential to detect small defects of the product. Our proposed model achieved 98% classification accuracy in the experiments and proved its efficiency on product inspection in real-time.

Enhancing Metal Surface Defect Recognition Through Image Patching and Synthetic Defect Generation

“…This step was taken to assess the impact of image patching and the SDGAN approaches, with the objective of selecting the most suitable dataset that can represent the diverse range of surface defects found in SMHs. Subsequently, we conducted an extensive comparative analysis by training popular classification models, including InceptionV3 [33], MobileNetV2 [34], EfficientNetV2S [35], ResNet50 [36], and SCA-CNN [37], alongside the baseline CNN and our proposed Inception-CNN model. These models were trained on the SMHSD-P-GAN dataset and evaluated using cross-validation.…”

“…By applying the inception module at different locations within the baseline CNN model, we aimed to evaluate its effectiveness and identify the optimal configuration for achieving improved classification accuracy and feature extraction. Additionally, we conducted a comparative analysis by fine-tuning well-known classification models, such as InceptionV3 [33], MobileNetV2 [34], EfficientNetV2S [35], ResNet50 [36], and CNN with SCA [37], and evaluated their performance on the test data.…”

YOLOX-Ray: An Efficient Attention-Based Single-Staged Object Detector Tailored for Industrial Inspections

“…By selectively attending to parts of an image, attention mechanisms can help models focus on relevant features, such as object boundaries or salient regions and ignore irrelevant information, such as background noise or occlusions. This can lead to improved accuracy and faster training times [ 13 , 14 ].…”