Convolutional Neural Networks (CNN) have become de facto state-of-the-art for the main computer vision tasks. However, due to the complex underlying structure their decisions are hard to understand which limits their use in some context of the industrial world. A common and hard to detect challenge in machine learning (ML) tasks is data bias. In this work, we present a systematic approach to uncover data bias by means of attribution maps. For this purpose, first an artificial dataset with a known bias is created and used to train intentionally biased CNNs. The networks' decisions are then inspected using attribution maps. Finally, meaningful metrics are used to measure the attribution maps' representativeness with respect to the known bias. The proposed study shows that some attribution map techniques highlight the presence of bias in the data better than others and metrics can support the identification of bias.
Weld quality inspection allows the detection of defects that may compromise the quality and strength of the weld. Although visual optical inspection offers lower reliability than other non-destructive methods, it enables weld analysis at a significantly lower cost. In this context, developing machine learning-based algorithms for automatic optical weld quality recognition requires acquiring large amounts of data for training. This entails high costs in terms of time, material and energy required for test preparation. However, one possible approach to tackling the problem with limited datasets is to use synthetic data. Using such data increases the amount and variety of data available to the detection algorithm. With a focus on the context of welding, this paper presents an approach that uses synthetic data as a form of data augmentation to improve the performance of the optical detection of weld seams. Specifically, we propose a generative neural network for semantic image synthesis using a limited starting dataset. The network generates new data instances by receiving as input a semantic map of the image to be represented. Weld defects such as porosity or weld spatter are added to the semantic map so that the network synthesizes corresponding defect images. Analysing the performance on a segmentation network, experimental results show how adding synthetic data to the original data can ensure improvements in network performance.
In sheet metal production the quality of a cut determines the conditions for a possible postprocessing. Considering the roughness as a parameter for assessing the quality of the cut edge, different techniques have been developed that use texture analysis and convolutional neural networks. All methods available require the use of appropriate equipment and work only in fixed light conditions. In order to discover new applications in the contexts of Industry 4.0, there is a necessity to go beyond their intrinsic limits as camera types and light condition while ensuring the same level of performance. Taking into account the strong increase of the smartphones features in recent years and the fact that their performance in some respect is now comparable to that of a PC with a middle-range mirrorless camera, it is no longer utopian to think of a new out-of-the-box use of these devices that employs the capability in a new way and in a new context. Therefore, we present a method that uses a mobile device with a camera to guarantee images of sufficient quality that can be used for further processing in order to determine the quality of the metal sheet edge. After the image acquisition of the sheet metal edge in real condition of use, the method uses a trained deep neural network to identify the sheet metal edge present in the picture. After the segmentation a no-reference image quality algorithm provides an image quality index, in terms of blurriness, for the image region of the cut edge. This way it is possible for the further evaluation of the cut edge to only consider image data that satisfies a specific quality, ignoring all the parts of the picture with a bad image quality.
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.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.