A Two-Stage Industrial Defect Detection Framework Based on Improved-YOLOv5 and Optimized-Inception-ResnetV2 Models

Aiming to address the currently low accuracy of domestic industrial defect detection, this paper proposes a Two-Stage Industrial Defect Detection Framework based on Improved-YOLOv5 and Optimized-Inception-ResnetV2, which completes positioning and classification tasks through two specific models. In order to make the first-stage recognition more effective at locating insignificant small defects with high similarity on the steel surface, we improve YOLOv5 from the backbone network, the feature scales of the feature fusion layer, and the multiscale detection layer. In order to enable second-stage recognition to better extract defect features and achieve accurate classification, we embed the convolutional block attention module (CBAM) attention mechanism module into the Inception-ResnetV2 model, then optimize the network architecture and loss function of the accurate model. Based on the Pascal Visual Object Classes 2007 (VOC2007) dataset, the public dataset NEU-DET, and the optimized dataset Enriched-NEU-DET, we conducted multiple sets of comparative experiments on the Improved-YOLOv5 and Inception-ResnetV2. The testing results show that the improvement is obvious. In order to verify the superiority and adaptability of the two-stage framework, we first test based on the Enriched-NEU-DET dataset, and further use AUBO-i5 robot, Intel RealSense D435 camera, and other industrial steel equipment to build actual industrial scenes. In experiments, a two-stage framework achieves the best performance of 83.3% mean average precision (mAP), evaluated on the Enriched-NEU-DET dataset, and 91.0% on our built industrial defect environment.

Semantic Graph Neural Network: A Conversion from Spam Email Classification to Graph Classification

In this study, we propose a method named Semantic Graph Neural Network (SGNN) to address the challenging task of email classification. This method converts the email classification problem into a graph classification problem by projecting email into a graph and applying the SGNN model for classification. The email features are generated from the semantic graph; hence, there is no need of embedding the words into a numerical vector representation. The method performance is tested on the different public datasets. Experiments in the public dataset show that the presented method achieves high accuracy in the email classification test against a few public datasets. The performance is better than the state-of-the-art deep learning-based method in terms of spam classification.

A Super-Resolution Reconstruction Driven Helmet Detection Workflow

The degrading of input images due to the engineering environment decreases the performance of helmet detection models so as to prevent their application in practice. To overcome this problem, we propose an end-to-end helmet monitoring system, which implements a super-resolution (SR) reconstruction driven helmet detection workflow to detect helmets for monitoring tasks. The monitoring system consists of two modules, the super-resolution reconstruction module and the detection module. The former implements the SR algorithm to produce high-resolution images, the latter performs the helmet detection. Validations are performed on both a public dataset as well as the realistic dataset obtained from a practical construction site. The results show that the proposed system achieves a promising performance and surpasses the competing methods. It will be a promising tool for construction monitoring and is easy to be extended to corresponding tasks.

Low Contrast Infrared Target Detection Method Based on Residual Thermal Backbone Network and Weighting Loss Function

Infrared (IR) target detection is an important technology in the field of remote sensing image application. The methods for IR image target detection are affected by many characteristics, such as poor texture information and low contrast. These characteristics bring great challenges to infrared target detection. To address the above problem, we propose a novel target detection method for IR images target detection in this paper. Our method is improved from two aspects: Firstly, we propose a novel residual thermal infrared network (ResTNet) as the backbone in our method, which is designed to improve the feature extraction ability for low contrast targets by Transformer structure. Secondly, we propose a contrast enhancement loss function (CTEL) that optimizes the weights about the loss value of the low contrast targets’ prediction results to improve the effect of learning low contrast targets and compensate for the gradient of the low-contrast targets in training back propagation. Experiments on FLIR-ADAS dataset and our remote sensing dataset show that our method is far superior to the state-of-the-art ones in detecting low-contrast targets of IR images. The mAP of the proposed method reaches 84% on the FLIR public dataset. This is the best precision in published papers. Compared with the baseline, the performance on low-contrast targets is improved by about 20%. In addition, the proposed method is state-of-the-art on the FLIR dataset and our dataset. The comparative experiments demonstrate that our method has strong robustness and competitiveness.

An IoT Crossdomain Access Decision-Making Method Based on Federated Learning

Crossdomain collaboration allows smart devices work together in different Internet of Things (IoT) domains. Trusted third party-based solutions require to fully understand the access information of the collaboration participants to implement crossdomain access control, which brings privacy risk. In this paper, we propose a federated learning-based crossdomain access decision-making method (FCAD), which builds a crossdomain access decision-making model without sharing privacy information of collaboration participants. Crossdomain access logs are extracted to construct a training dataset. Data enhancement method is used to address the uneven distribution of the dataset. Federated learning and gradient aggregation methods are used to prevent privacy leaks. The experiments on the public dataset show that FCAD obtains a prediction accuracy of 83.6% in the existing crossdomain access system.

An Improved 2D U-Net Model Integrated Squeeze-and-Excitation Layer for Prostate Cancer Segmentation

In this paper, we proposed an improved 2D U-Net model integrated squeeze-and-excitation layer for prostate cancer segmentation. The proposed model combined a more complex 2D U-Net model and squeeze-and-excitation technique. The model consisted of an encoder stage and a decoder stage. The encoder stage aims to extract features of the input, which contains CONV blocks, SE layers, and max-pooling layers for improving the feature extraction capability of the model. The decoder aims to map the extracted features to the original image with CONV blocks, SE layers, and upsampling layers. The SE layer is implemented to learn more global and local features. Experiments on the public dataset PROMISE12 have demonstrated that the proposed model could achieve state-of-the-art segmentation performance compared with other traditional methods.

Continuous Viewpoint Planning in Conjunction with Dynamic Exploration for Active Object Recognition

Active object recognition (AOR) aims at collecting additional information to improve recognition performance by purposefully adjusting the viewpoint of an agent. How to determine the next best viewpoint of the agent, i.e., viewpoint planning (VP), is a research focus. Most existing VP methods perform viewpoint exploration in the discrete viewpoint space, which have to sample viewpoint space and may bring in significant quantization error. To address this challenge, a continuous VP approach for AOR based on reinforcement learning is proposed. Specifically, we use two separate neural networks to model the VP policy as a parameterized Gaussian distribution and resort the proximal policy optimization framework to learn the policy. Furthermore, an adaptive entropy regularization based dynamic exploration scheme is presented to automatically adjust the viewpoint exploration ability in the learning process. To the end, experimental results on the public dataset GERMS well demonstrate the superiority of our proposed VP method.

Imaginary Finger Movements Decoding Using Empirical Mode Decomposition and a Stacked BiLSTM Architecture

Motor Imagery Electroencephalogram (MI-EEG) signals are widely used in Brain-Computer Interfaces (BCI). MI-EEG signals of large limbs movements have been explored in recent researches because they deliver relevant classification rates for BCI systems. However, smaller and noisy signals corresponding to hand-finger imagined movements are less frequently used because they are difficult to classify. This study proposes a method for decoding finger imagined movements of the right hand. For this purpose, MI-EEG signals from C3, Cz, P3, and Pz sensors were carefully selected to be processed in the proposed framework. Therefore, a method based on Empirical Mode Decomposition (EMD) is used to tackle the problem of noisy signals. At the same time, the sequence classification is performed by a stacked Bidirectional Long Short-Term Memory (BiLSTM) network. The proposed method was evaluated using k-fold cross-validation on a public dataset, obtaining an accuracy of 82.26%.

Klasifikasi Jenis Ikan Laut Menggunakan Metode SVM dengan Fitur HOG dan HSV

Ikan merupakan hewan vertebrata yang hidup di dalam air. Ikan memiliki insang yang berfungsi sebagai alat pernapasan untuk mengambil oksigen dalam air dan sirip digunakan untuk berenang. Dalam kelompok hewan vertebrata, ikan memiliki jumlah terbesar diperkirakan 40.000 spesies, sementara yang tercatat hingga saat ini sekitar 25.000. Terdapat sekitar 13.630 spesies ikan yang berada di perairan laut, dikarenakan hampir 70% permukaan bumi terdiri dari air laut dan perairan tawar hanya sekitar 1% saja. Penelitian ini menggunakan citra ikan laut yang diambil dari public dataset yang memiliki 7 jenis ikan laut dimana setiap jenis ikan laut ada 7.000 gambar yang akan dilakukan tahap segmentasi warna HSV dengan mengambil nilai value sehingga menjadi grayscale yang akan dilanjutkan ke proses HOG dan untuk mengklasifikasikan jenis ikan laut menggunakan SVM. Untuk teknik pengujian dan pembagian dataset menggunakan metode Fold Cross Validation jenis Leave One Out (LOO). Berdasarkan hasil pengujian klasifikasi SVM baik kernel linear maupun polynomial dengan menggunakan 3-Fold, 4-Fold, dan 5-Fold. Accuracy tertinggi jenis ikan Black Sea Sprat senilai 94,06%. Untuk jenis ikan Gilt Head Bream tertinggi didapat senilai 94,31%. Selanjutnya jenis ikan Hourse Mackerel mendapatkan nilai accuracy tertinggi 94,74%. Kemudian jenis ikan Red Mullet nilai accuracy tertinggi sebesar 94,76%. Selanjutnya jenis ikan Red Sea Bream memperoleh nilai accuracy tertinggi 94,86%, jenis ikan Sea Bass yang dengan nilai accuracy tertinggi 77,86% dan ikan Striped Red Mullet memperoleh nilai accuracy tertinggi 94,41%.

A Deep Adversarial Model for Segmentation Assisted COVID-19 Diagnosis Using CT Images

The outbreak of coronavirus disease 2019(COVID-19) is spreading rapidly around the world, resulting in a global pandemic. Imaging techniques such as computed tomography (CT) play an essential role in the diagnosis and treatment of the disease since lung infection or pneumonia is a common complication. However, training a deep network to learn how to diagnose COVID-19 rapidly and accurately in CT images and segment the infected regions like a radiologist is challenging. Since the infectious area are difficult to distinguish, and manually annotation the segmentation results is time-consuming. To tackle these problems, we propose an efficient method based on a deep adversarial network to segment the infection regions automatically. Then the predicted segment results can assist the diagnosis network in identifying the COVID-19 samples from the CT images. On the other hand, a radiologist-like segmentation network provides detailed information of the infectious regions by separating areas of ground-glass, consolidation, and pleural effusion, respectively. Our method can accurately predict the COVID-19 infectious probability and provides lesion regions in CT images with limited training data. Additionally, we have established a public dataset for multitask learning. Extensive experiments on diagnoses and segmentation show superior performance over state-of-the-art methods.