Road Curb Detection Based on a Deep Learning Framework

Fault feature extraction is critical for fault line detection, but difficult to be effective and robust. Unbalanced characteristics of the fault signal sample will make feature extraction more difficult. A novel method using Choi- Williams time-frequency distribution based convolutional neural network and task decomposition framework was proposed. Choi- Williams time-frequency analysis was applied to generate time-frequency distribution image of fault signal. Then, convolutional neural network (CNN) was trained by a lot of time-frequency distribution images generated under different fault conditions. CNN can extract features of the time-frequency distribution image adaptively and select the fault line. The task decomposition framework was first proposed to solve the problem of unbalanced fault signal sample for better feature extraction. A resonant grounding distribution system is simulated to verify this method under different fault conditions and the results showed the detection of the single line-to-ground fault is more accurate.

Aiming at the problems of slow convergence and low recognition accuracy of the existing image big data recognition methods, a convolution neural network based image big data recognition method is proposed. In the concept of machine vision, image recognition refers to the ability of software to distinguish characters, positions, objects, actions and handwriting in pictures. Computers can apply machine vision techniques, combined with artificial intelligence and cameras to image recognition. Describes the convolution layer, pool layer, full connection layer and classification layer in the model. MNIST data sets are simulated to verify the superiority of the proposed method. Convolutional neural network has achieved good results in the field of image recognition, but its network structure has a great impact on the effect and efficiency of image recognition. In order to improve the recognition performance, a new convolution neural network structure is designed and implemented by repeated use of smaller convolution kernels. The results show that compared with other similar algorithms, the false recognition rate of this method is low, and the lower average square error of classification can be realized in a short time. This paper expounds the basic flow of digital recognition system from the aspects of data extraction and preprocessing, feature extraction and selection, classifier design and so on[1].

This paper introduces a predictive model for Chronic Kidney Disease (CKD) classification with the application of deep learning through Convolutional Neural Networks (CNNs). The suggested system utilizes CNNs as feature extractors with layers including convolutional, global average pooling, dropout, and dense layers for classification. The model is trained with a vast clinical dataset with patient records and lab results, with the goal of accurately differentiating between CKD and non-CKD cases. The model is trained with the Adam optimizer with a training period of over 50 epochs, with data augmentation strategies such as normalization and one-hot encoding providing it with added strength. Evaluation metrics such as accuracy, precision, recall, and F1 score are used in selecting the best-performing model. Practical usability is ensured by the application of a Flask-based web application for real-time data input and risk estimation. Visualizations such as confusion matrices and learning curves give insight into the performance of the model as well as decision making processes. The predictive system presents substantial implications for the detection of CKD at an early stage, enabling proactive treatment approaches as well as enhancing patient outcomes.

4D digital tomosynthesis (DTS) techniques for image-guided radiation therapy (IGRT) are able to reduce radiation dose, scan and reconstruction time compared to 4D cone-beam computed tomography (CBCT). In spite of these benefits, the 4D DTS techniques cause the degradation of image quality due to an intrinsic imaging strategy and consequently reduce treatment accuracy. In this study, a deep learning-based convolutional neural network (CNN) framework was proposed for 4D DTS imaging. The proposed CNN framework consisted of the data restoration network based on a U-Net and the denoising network combined with a 2D wavelet transform, and the network training was implemented with clinical images. The quality of the 4D DTS images obtained from the proposed model was evaluated in terms of quantitative accuracy, spatial resolution and noise property. The results showed that the proposed CNN framework improved the quantitative accuracy of 4D DTS images by 3–19%, and the spatial resolution and noise for the proposed CNN framework were reduced by 2.24–7.33% and 8.92–40.07%, respectively, in comparison to other imaging models. These results represented that the degradation of the 4D DTS image quality can be recovered using the proposed CNN framework, and the proposed model is suitable for maintaining spatial resolution as well as suppressing noise and artifacts. In conclusion, the proposed CNN framework can be potentially used to improve the quality of 4D DTS images for the IGRT.

To investigate the recognition effect of flag motions based on 9 axis sensors, starting with deep learning methods, this paper proposes a framework for feature extraction and recognition of flag movements recognition by employing an improved Inception-ResNet dual-stream network. In traditional signal recognition studies, Support Vector Machine (SVM), Random Forest (RF) and One Dimensional Convolutional Neural Network (CNN) are usually used to extract signals' features. In the meanwhile, the time series data sets such as flag movements are usually the standard data set after processing. Therefore, there are usually some limitations in traditional systems. First, in actual environment, there exists a lack of effective segmentation detection method for the samples of long time series, resulting in the deviation of the data set in the recognition process. Second, the One-Dimensional CNN framework used and the machine learning frameworks used in previous studies are difficult to process large quantities of data with too high computational memory. Based on these problems, this study proposes a signal change point detection model based on the diversity factory function in the signal segmentation and detection stage, miniaturizes the convolution kernel in the original CNN by using the Inception-ResNet(I-R) dual-flow network separable convolution method, and proposes a CrossEntropy-Logistic(C-L) joint classification loss function. Through conducting comparative experiments, it is found that the average calculation parameter of the CNN framework based on the Inception-ResNet model is 2.7×10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">7</sup> , which is approximately 37% lower than the number of 3.7×10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">7</sup> in the original CNN model. Finally, the recognition rate between C-L joint loss function and other models such as Multi-Layer Perceptron and Ensemble Learning in recent years are compared. Compared with Ensemble Learning-CrossEntropy (ELC) model, the C-L joint loss function can improve the recognition rate by nearly 5% according to the results of flag movements identification measured by several classification models.

Fire outbreaks continue to cause damage despite the improvements in fire-detection tools and algorithms. As the human population and global warming continue to rise, fires have emerged as a significant worldwide issue. These factors may contribute to the greenhouse effect and climatic changes, among other detrimental consequences. It is still challenging to implement a well-performing and optimized approach, which is sufficiently accurate, and has tractable complexity and a low false alarm rate. A small fire and the identification of a fire from a long distance are also challenges in previously proposed techniques. In this study, we propose a novel hybrid model, called IS-CNN-LSTM, based on convolutional neural networks (CNN) to detect and analyze fire intensity. A total of 21 convolutional layers, 24 rectified linear unit (ReLU) layers, 6 pooling layers, 3 fully connected layers, 2 dropout layers, and a softmax layer are included in the proposed 57-layer CNN model. Our proposed model performs instance segmentation to distinguish between fire and non-fire events. To reduce the intricacy of the proposed model, we also propose a key-frame extraction algorithm. The proposed model uses Internet of Things (IoT) devices to alert the relevant person by calculating the severity of the fire. Our proposed model is tested on a publicly available dataset having fire and normal videos. The achievement of 95.25% classification accuracy, 0.09% false positive rate (FPR), 0.65% false negative rate (FNR), and a prediction time of 0.08 s validates the proposed system.

During the past five years, there has been an increasing trend of weakly supervised crowd counting methods being developed since such methods just rely on count-level annotations and avoid a laborious labeling process. But, the existing weakly supervised methods usually fail to achieve comparable counting performance to the fully supervised methods. To improve the accuracy of crowd counting tasks, we propose to combine the convolutional neural network (CNN) and Transformer frameworks. Since CNN focuses on capturing local detail information and Transformer can effectively extract global context information, we believe that the combination of CNN and Transformer could learn more efficient feature representations for crowd images. Our proposed framework is named CrowdCCT (Crowd Counting via CNN and Transformer), and it is composed of a CNN feature extraction part, a Transformer feature extraction part, and a counting regression part. In the CNN part, we utilize DenseNet121 to learn rich semantic features with its inherent dense connection structure. In the Transformer part, we introduce two attention modules, Multi-Scale Dilated Attention (MSDA) and Location-Enhanced Attention (LEA), working together to extract more expressive features. The output features are then fed into the regression part to generate the predicted counting results. Experiments on four crowd counting benchmark datasets demonstrate that our proposed CrowdCCT can achieve superior performance. Also, the experimental results validate the feasibility and effectiveness of combining CNN and Transformer for weakly supervised counting tasks. Our work could be expected to promote further combination research on CNN and Transformer.

In traditional 3D reconstruction methods, using a single view to predict the 3D structure of an object is a very difficult task. This research mainly discusses human pose recognition and estimation based on 3D multiview basketball sports dataset. The convolutional neural network framework used in this research is VGG11, and the basketball dataset Image Net is used for pretraining. This research uses some modules of the VGG11 network. For different feature fusion methods, different modules of the VGG11 network are used as the feature extraction network. In order to be efficient in computing and processing, the multilayer perceptron in the network model is implemented by a one‐dimensional convolutional network. The input is a randomly sampled point set, and after a layer of perceptron, it outputs a feature set of n × 16. Then, the feature set is sent to two network branches, one is to continue to use the perceptron method to generate the feature set of n × 1024, and the other network is used to extract the local features of points. After the RGB basketball sports picture passes through the semantic segmentation network, a picture containing the target object is obtained, and the picture is input to the constructed feature fusion network model. After feature extraction is performed on the RGB image and the depth image, respectively, the RGB feature, the local feature of the point cloud, and the global feature are spliced and fused to form a feature vector of N × 1152. There are three branches for this vector network, which, respectively, predict the object position, rotation, and confidence. Among them, the feature dimensionality reduction is realized by one‐dimensional convolution, and the activation function is the ReLU function. After removing the feature mapping module, the accuracy of VC‐CNN_v1 dropped by 0.33% and the accuracy of VC‐CNN_v2 dropped by 0.55%. It can be seen from the research results that the addition of the feature mapping module improves the recognition effect of the network to a certain extent

In this study, a new method for the analysis of harmonic components in the power system based on a deep learning (DL) framework is introduced. In the proposed method, both amplitudes and phases of the harmonic components can be estimated accurately, unlike most of the research work in the literature, which usually focus on estimating amplitudes only. A convolutional neural network (CNN) structure is used to estimate the phases and amplitudes of harmonics, although CNN is usually used for classification. It has been shown that the proposed DL‐based method can satisfactorily estimate both amplitudes and phases of the power system harmonics inside a 20‐ms window and this makes the proposed method suitable for possible real‐time applications, such as active power filtering of the harmonics. It has also been shown that the proposed method is robust to fundamental frequency changes. Experiments on carefully‐generated data sets to reflect the power system behaviour show that the proposed method demonstrates remarkably good performance in terms of estimation accuracy, especially for time‐varying frequency cases. Average error for the amplitude estimation is obtained as 0.21% and that for the phase is 9°, which outperforms the other compared analyses methods in cases of fundamental frequency variations.

Convolutional autoencoders based image denoising has emerged as a promising approach in medical imaging applications. Most of the traditional denoising methods such as Bayesian filter, non-linear median filter, wavelet based shearlet transform etc, are difficult to process medical images due to the existence of additive, multiplicative and Gaussian noise. Also, these models cannot resolve the issue of sparsity in the compressed medical images. To overcome these issues, a novelcompressive sensing based multilevel denoising technique is proposed to minimize the noise rate of medical images in the Convolution neural network(CNN) framework. This technique is implemented in convolutional autoencoders framework to reduce the noise in each deep network layer and to improve the quality of the images. This technique efficiently denoise the medical images using the inter, intra variance in the CNN framework. Experimental results demonstrated that the proposed framework has better PSNR and SSIM measures compared to the traditional state-of-the-art approaches.

Accurate and early diagnosis of malaria from peripheral blood smear images remains a critical challenge in healthcare, particularly in resource-limited settings. In this work, we propose an optimized convolutional neural network (CNN) framework enhanced by Otsu thresholding-based image segmentation for improved detection of malaria-infected cells. A dataset of 43,400 blood smear images was utilized, divided into a 70:30 ratio for training and testing. A baseline 12-layer CNN achieved 95% accuracy, which improved to 97% with the integration of EfficientNet-B7 through a hybrid parallel feature-fusion model. Further enhancement was achieved using Otsu-based segmentation, where preprocessing emphasized parasite-relevant regions while retaining morphological context in the RGB images. This approach yielded the highest accuracy of 97.96%, reflecting a ~ 3% gain over the baseline CNN. To ensure the reliability of the segmentation step, we created a manually annotated subset of 100 images and computed quantitative segmentation metrics by comparing Otsu-generated masks with reference masks. The method achieved a mean Dice coefficient of 0.848 and Jaccard Index (IoU) of 0.738, confirming that Otsu segmentation effectively isolates parasitic regions despite its simplicity. Five-fold cross-validation was also performed, yielding consistent results (94.8%, 96.9%, and 97.8%), thereby supporting the robustness of the framework. The proposed pipeline demonstrates that simple yet effective preprocessing can significantly boost CNN-based classification while maintaining interpretability and computational feasibility. These findings suggest that segmentation-driven deep learning frameworks can play a vital role in developing reliable, scalable, and cost-effective malaria diagnostic tools.

Non-syndromic craniosynostosis (NSCS) is a disease, in which a single cranial bone suture is prematurely fused. The early intervention of the disease is associated with a favorable outcome at a later age, so appropriate screening of NSCS is essential for its clinical management. The present study aims to develop a classification and detection system of NSCS using skull X-ray images and a convolutional neural network (CNN) deep learning framework. A total of 56 NSCS cases (scaphocephaly [ n = 17], trigonocephaly [n = 28], anterior plagiocephaly [n = 8], and posterior plagiocephaly [n = 3]) and 25 healthy control infants were included in the study. All the cases underwent skull X-rays and computed tomography scan for diagnosis in our institution. The lateral views obtained from the patients were retrospectively examined using a CNN framework. Our CNN model classified the 4 NSCS types and control with high accuracy (100%). All the cases were correctly classified. The proposed CNN model may offer a safe and high-sensitivity screening of NSCS and facilitate early diagnosis of the disease and better neurocognitive outcome for patients.

Millions of images and videos are daily being shared on internet due to recent advances in social media and multimedia technology. However, many of them are manipulated using readily available image and video editing software and apps. Specifically, “DeepFake” videos, which are created by swapping an individual’s face with the face of other individual by utilising deep learning techniques, may constitute a substantial risk either to attack character of public figures or to fool automated face recognition systems. In this paper, a hybrid and high-confidence DeepFake face video detection framework is proposed to discriminate manipulated face videos from non-manipulated face videos. The proposed framework is composed of three main stages: face detection, deep feature extraction, and long short-term memory (LSTM) classification. The face detection is achieved by a well-known Viola–Jones detector. The detected face images are rescaled to make them compatible with an input of the pre-trained convolutional neural network model. The residual network with 50 layers is employed for deep feature extraction. Finally, the LSTM model is considered in classification to gather final verdict: bona fide or DeepFake. The considered LSTM network contains seven layers such as input layer, two biLSTM layers, dropout layer, fully connected layer, softmax layer, and classification layer. The proposed framework is capable of spotting DeepFake face videos with high accuracy that is substantiated via experimental analysis on two diverse and publicly available datasets (i.e., Celeb-DF and DeepFakeTIMIT).

Diabetes is a chronic disease, if not treated in time may lead to many complications including diabetic foot ulcers (DFU). DFU is a dangerous disease, it needs regular treatment otherwise it may lead towards foot amputation. The DFU is classified into two categories such as infection (bacteria) and the ischaemia (inadequate supply of the blood). The DFU detection at an initial phase is a tough procedure. Therefore in this research work 16 layers convolutional neural network (CNN) for example 01 input, 03 convolutional, 03 batch-normalization, 01 average pooling, 01 skips convolutional, 03 ReLU, 01 add (element-wise addition of two inputs), fully connected, softmax and classification output layers for classification and YOLOv2-DFU for localization of infection/ischaemia models are proposed. In the classification phase, deep features are extracted and supplied to the number of classifiers such as KNN, DT, Ensemble, softmax, and NB to analyze the classification results for the selection of best classifiers. After the experimentation, we observed that DT and softmax achieved consistent results for the detection of ischaemia/infection in all performance metrics such as sensitivity, specificity, and accuracy as compared with other classifiers. In addition, after the classification, the Gradient-weighted class activation mapping (Grad-Cam) model is used to visualize the high-level features of the infected region for better understanding. The classified images are passed to the YOLOv2-DFU network for infected region localization. The Shuffle network is utilized as a mainstay of the YOLOv2 model in which bottleneck extracted features through ReLU node-199 layer and passed to the YOLOv2 model. The proposed method is validated on the newly developed DFU-Part (B) dataset and the results are compared with the latest published work using the same dataset.

To distinguish benign from malignant pulmonary nodules using CT images is critical for their precise diagnosis and treatment. A new Agile convolutional neural network (CNN) framework is proposed to conquer the challenges of a small-scale medical image database and the small size of the nodules, and it improves the performance of pulmonary nodule classification using CT images. A hybrid CNN of LeNet and AlexNet is constructed through combining the layer settings of LeNet and the parameter settings of AlexNet. A dataset with 743 CT image nodule samples is built up based on the 1018 CT scans of LIDC to train and evaluate the Agile CNN model. Through adjusting the parameters of the kernel size, learning rate, and other factors, the effect of these parameters on the performance of the CNN model is investigated, and an optimized setting of the CNN is obtained finally. After finely optimizing the settings of the CNN, the estimation accuracy and the area under the curve can reach 0.822 and 0.877, respectively. The accuracy of the CNN is significantly dependent on the kernel size, learning rate, training batch size, dropout, and weight initializations. The best performance is achieved when the kernel size is set to [Formula: see text], the learning rate is 0.005, the batch size is 32, and dropout and Gaussian initialization are used. This competitive performance demonstrates that our proposed CNN framework and the optimization strategy of the CNN parameters are suitable for pulmonary nodule classification characterized by small medical datasets and small targets. The classification model might help diagnose and treat pulmonary nodules effectively.