Quantitative Analysis of Metallographic Image Using Attention-Aware Deep Neural Networks

In recent years, with the development of computer image processing technology and quantitative metallography, quantitative metallographic analysis and image restoration based on machine vision image processing have attracted extensive attention. To deeply study the special ability of machine vision technology in metallographic image restoration and quantitative analysis, we adopt a model construction method, parameter design method and comparative analysis method to collect samples and analyze the machine vision system. Through quantitative metallographic organization, based on the progress in artificial neural networks and classification algorithms, a set of analysis system equipment and image restoration algorithm was finally established. The experimental results show that the recognition rate of this method is 5.69% higher than that of method 3, and the overall recognition rate is 5.1% higher than that of method 3, reaching 98.65%. It basically proves the superiority of the image classification algorithm in this paper and also indirectly proves that the system can play a role in the restoration and quantitative analysis of metallographic tissue images.

Grain size evaluation is a key factor in assessing the mechanical properties and service performance of stainless steel. However, conventional manual analysis of metallographic images is time-consuming, subjective, and often lacks reproducibility, limiting its industrial applicability. To overcome these challenges, this paper proposes CS-SegFormer (Channel and Spatial), an improved semantic segmentation framework tailored for metallographic image analysis of austenitic stainless steel. Built upon SegFormer, the proposed model integrates channel and spatial attention modules, which enhance the perception of grain boundaries, fine-scale textures, and complex microstructural variations.Experiments were conducted on a self-constructed dataset containing 120 ultra-high-resolution metallographic images, annotated into six categories including grains, twins, precipitates, and defects. Comparative results demonstrate that CS-SegFormer outperforms representative segmentation networks—UNet, KNet, SegFormer, and FastSCNN—achieving improvements of approximately 10% in both mIoU and mDice. These findings validate the superior accuracy and robustness of CSSegFormer and highlight its potential for reliable automated grain size grading in metallographic evaluation and industrial applications.

An investigation has been made on the infiltration behaviour in rapid prototyping of Al alloy parts under different infiltration atmospheres. An Al 6061 alloy preform was first prepared by selective laser sintering. Then the aluminium precursor powders in the preform were converted into AlN to form a skeletal AlN structure, which was subsequently pressureless infiltrated with a molten Al 6061 alloy under nitrogen, argon or vacuum. The pathway of the infiltrant was determined by density measurements in conjunction with metallographic examination and quantitative image analysis. Detailed comparison of the microstructures at the surface and the centre of the as-infiltrated samples indicates that the pathway into the porous preform is dependent on the infiltration atmosphere. Under vacuum, the infiltrant fills the interior of the preform first and then propagates to the surface. In contrast, under nitrogen or argon, the infiltrant penetrates along the surface prior to filling the interior.

Training very deep neural networks: Rethinking the role of skip connections

Classification of graphite particles in metallographic images of cast irons – Quantitative image analysis versus deep learning

Stroke is a kind of cerebrovascular disease that heavily damages people’s life and health. The quantitative analysis of brain MRI images plays an important role in the diagnosis and treatment of stroke. Deep neural networks with massive data learning ability supply a powerful tool for lesion detection. In order to study the property of the stroke lesions and complete intelligent automatic detection, we collaborated with two authoritative hospitals and collected 5,668 brain MRI images of 300 ischemic stroke patients. All the lesion regions in the images were accurately labeled by professional doctors to ensure the authority and effectiveness of the data. Three categories of deep learning object detection networks including Faster R-CNN, YOLOV3, and SSD are applied to implement automatic lesion detection with the best precision of 89.77%. Meanwhile, statistical analysis of the locations, shapes of the lesions, and possible related diseases is conducted with valid conclusions. The research contributes to the intelligent assisted diagnosis and prevention and treatment of ischemic stroke.

Automatic recognition of surface defects for hot-rolled steel strip based on deep attention residual convolutional neural network

As an enabling technology for emerging and future generations of wireless networks, sparse code multiple access (SCMA) offers major improvements in terms of spectral efficiency and massive connectivity. Although the message passing algorithm (MPA) for SCMA decoding at the receiver side can achieve near optimum performance, it entails high computational complexity. In this paper, to address this issue, we propose a novel SCMA decoder based on deep residual neural network (ResNet), wherein the decoder is trained to predict the transmit codewords. In our approach, residual blocks are employed to tackle the problems of accuracy saturation and vanishing gradients with deep learning based decoder, while batch normalization is utilized to enhance the stability and robustness of the decoder. The performance of the proposed ResNet decoder for SCMA is validated by means of simulations over AWGN and Rayleigh fading channels. The results show that besides a much reduced complexity, the proposed decoder leads to improvements in term of bit error rate (BER) over competing deep neural network (DNN) based decoders.

Gait recognition is a noncontact biometric procedure that determines the identity or health status of a person by analyzing his or her walking posture and habits, including skeletal and joint movements. The most remarkable feature of this method is the possibility of conducting recognition without demanding much cooperation from participants. Therefore, this recognition technique has attracted much attention from scholars. Additionally, because of the rapid development of graphics processing unit technology, related hardware and computation performance, the applications of deep-learning technology are considerably enhanced. The objective of this study was to apply a deep neural network (DNN), which employs deep-learning technology, to achieve gait-based automatic pedestrian detection and recognition. In contrast to using wearable devices to precisely capture skeletal and joint movements, pedestrian color-image sequences were used as input in this study. Subsequently, a pretraining convolutional neural network (CNN) was employed to capture pedestrian location and extract pedestrian dense optical flow to serve as concrete low-level feature inputs. Then, a finely-tuned DNN based on the wide residual network was employed to extract high-level abstract features. In addition, to overcome the difficulty of obtaining local temporal features by using a 2D CNN, part of the 3D convolutional structure was introduced into the CNN. This design enabled use of limited memory to acquire more effective features and enhance the DNN performance. The experimental results show that the proposed method has exceptional performance for pedestrian detection and recognition.

Alpha-methylacyl-CoA racemase (AMACR) is an enzyme involved in the metabolism of fatty acids and is an important tissue biomarker in the prostate to distinguish normal glands from prostate cancer. Here, for the first time, we evaluated the expression of AMACR protein in normal breast, ductal carcinoma in situ, and invasive carcinomas. By immunofluorescence and immunohistochemistry, AMACR was seen in cytoplasmic granules consistent with a mitochondrial and peroxisomal localization. AMACR expression was determined by immunohistochemistry on 160 invasive carcinomas with long follow-up, using a high-density tissue microarray, and evaluated by two methods: standard pathology review and quantitative image analysis. AMACR was overexpressed in 42 of 160 (26%) invasive carcinomas, and it was associated with a decrease in tumor differentiation, a feature of aggressive breast cancer. Quantitative analysis allowed for better discrimination and more accurate evaluation of low-intensity staining. In conclusion, AMACR protein is expressed in normal breast and its expression seems to increase in invasive carcinomas. We observed stronger AMACR protein expression in high-grade carcinomas when compared with low-grade ones. Quantitative image analysis is a novel way to accurately and reproducibly evaluate immunohistochemistry in breast tissue samples using high-density tissue microarrays.

Archaeopolymetallurgical study of materials from an Iberian culture site in Spain by scanning electron microscopy with X-ray microanalysis, chemometrics and image analysis

Advancements in genomic, computing, and imaging technology have spurred new opportunities to use quantitative image analysis (QIA) for diagnostic testing. To develop evidence-based recommendations to improve accuracy, precision, and reproducibility in the interpretation of human epidermal growth factor receptor 2 (HER2) immunohistochemistry (IHC) for breast cancer where QIA is used. The College of American Pathologists (CAP) convened a panel of pathologists, histotechnologists, and computer scientists with expertise in image analysis, immunohistochemistry, quality management, and breast pathology to develop recommendations for QIA of HER2 IHC in breast cancer. A systematic review of the literature was conducted to address 5 key questions. Final recommendations were derived from strength of evidence, open comment feedback, expert panel consensus, and advisory panel review. Eleven recommendations were drafted: 7 based on CAP laboratory accreditation requirements and 4 based on expert consensus opinions. A 3-week open comment period received 180 comments from more than 150 participants. To improve accurate, precise, and reproducible interpretation of HER2 IHC results for breast cancer, QIA and procedures must be validated before implementation, followed by regular maintenance and ongoing evaluation of quality control and quality assurance. HER2 QIA performance, interpretation, and reporting should be supervised by pathologists with expertise in QIA.

Simply feeding of a last hidden layer of the deep neural network (DNN) back to the input layer recently found to be effective for noise robust acoustic modeling. Such high level feature strengthens the robustness of DNN based acoustic model while paying approximately twice the computational cost. In this paper, we proposed to feed such high level feature iteratively back to lower layers, which is referred as multi-scale feedback connection. With this intention, we firstly extract the high level feature at the last hidden layer of DNN. Second, this high level feature feed back to a lower scale features, they then generates a subsequent prediction as well as a subsequent high level feature. This subsequent high level feature is further feed down to a lower layers. We evaluated the proposed approach on both TIMIT and a large scale internal dataset. The large scale internal dataset includes voice search and far field dataset. Our finding is two aspects. First, at equivalent computational costs, the multiscale feedback connection outperforms the DNN, the DNN with skip connection and the DNN with feedback connection. The improvement is larger on the far field dataset. Second, pair layers-wise pretraining helps the proposed approach to converge better.

Annotation of foot-contact and foot-off events is the initial step in post-processing for most quantitative gait analysis workflows. If clean force plate strikes are present, the events can be automatically detected. Otherwise, annotation of gait events is performed manually, since reliable automatic tools are not available. Automatic annotation methods have been proposed for normal gait, but are usually based on heuristics of the coordinates and velocities of motion capture markers placed on the feet. These heuristics do not generalize to pathological gait due to greater variability in kinematics and anatomy of patients, as well as the presence of assistive devices. In this paper, we use a data-driven approach to predict foot-contact and foot-off events from kinematic and marker time series in children with normal and pathological gait. Through analysis of 9092 gait cycle measurements we build a predictive model using Long Short-Term Memory (LSTM) artificial neural networks. The best-performing model identifies foot-contact and foot-off events with an average error of 10 and 13 milliseconds respectively, outperforming popular heuristic-based approaches. We conclude that the accuracy of our approach is sufficient for most clinical and research applications in the pediatric population. Moreover, the LSTM architecture enables real-time predictions, enabling applications for real-time control of active assistive devices, orthoses, or prostheses. We provide the model, usage examples, and the training code in an open-source package.

Lightweight single image super-resolution with attentive residual refinement network