In order to improve the detection accuracy of the surface defect detection of industrial hot rolled strip steel, the advanced technology of deep learning is applied to the surface defect detection of strip steel. In this paper, we propose a framework for strip surface defect detection based on a convolutional neural network (CNN). In particular, we propose a novel multi-scale feature fusion module (ATPF) for integrating multi-scale features and adaptively assigning weights to each feature. This module can extract semantic information at different scales more fully. At the same time, based on this module, we build a deep learning network, CG-Net, that is suitable for strip surface defect detection. The test results showed that it achieved an average accuracy of 75.9 percent (mAP50) in 6.5 giga floating-point operation (GFLOPs) and 105 frames per second (FPS). The detection accuracy improved by 6.3% over the baseline YOLOv5s. Compared with YOLOv5s, the reference quantity and calculation amount were reduced by 67% and 59.5%, respectively. At the same time, we also verify that our model exhibits good generalization performance on the NEU-CLS dataset.
Integrated circuit (IC) X-ray wire bonding image inspections are crucial for ensuring the quality of packaged products. However, detecting defects in IC chips can be challenging due to the slow defect detection speed and the high energy consumption of the available models. In this paper, we propose a new convolutional neural network (CNN)-based framework for detecting wire bonding defects in IC chip images. This framework incorporates a Spatial Convolution Attention (SCA) module to integrate multi-scale features and assign adaptive weights to each feature source. We also designed a lightweight network, called the Light and Mobile Network (LMNet), using the SCA module to enhance the framework’s practicality in the industry. The experimental results demonstrate that the LMNet achieves a satisfactory balance between performance and consumption. Specifically, the network achieved a mean average precision (mAP50) of 99.2, with 1.5 giga floating-point operations (GFLOPs) and 108.7 frames per second (FPS), in wire bonding defect detection.
In 3D packaging memory devices, solder joints are critical links between the chip and the printed circuit board (PCB). Under severe working conditions, cracks inevitably occur due to thermal shock. If cracks grow in the solder joint, the chip will be disconnected with the PCB, causing its function failure. In this paper, the reliability of solder joints under thermal shock are evaluated for 3D packaging memory devices by means of the SEM and finite element analysis. As microscopically studied by the SEM, it is found out that the main failure mechanism of solder joints in such test is the thermal fatigue failure of solder joints. Finite element analysis shows that cracks are caused by the accumulation of plastic work and creep strain. The initiation and growth of cracks are mainly influenced by the inelastic strain accumulation. The trends of cracks are influenced by the difference between the coefficient of thermal expansion (CTE) of epoxy resin and that of the chip.
In order to improve the production quality and qualification rate of chips, X-ray nondestructive imaging technology has been widely used in the detection of chip defects, which represents an important part of the quality inspection of products after packaging. However, the current traditional defect detection algorithm cannot meet the demands of high accuracy, fast speed, and real-time chip defect detection in industrial production. Therefore, this paper proposes a new multi-scale feature fusion module (ATSPPF) based on convolutional neural networks, which can more fully extract semantic information at different scales. In addition, based on this module, we design a deep learning model (ATNet) for detecting lead defects in chips. The experimental results show that at 8.2 giga floating point operations (GFLOPs) and 146 frames per second (FPS), mAP0.5 and mAP0.5–0.95 can achieve an average accuracy of 99.4% and 69.3%, respectively, while the detection speed is faster than the baseline yolov5s by nearly 50%.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2025 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.