One of the most promising techniques used in various sciences is deep neural networks (DNNs). A special type of DNN called a convolutional neural network (CNN) consists of several convolutional layers, each preceded by an activation function and a pooling layer. The feature map of the previous layer is sampled by the pooling layer (that seems to be an important layer) to create a new feature map with condensed resolution. This layer significantly reduces the spatial dimension of the input. It always accomplished two main goals. As a first step, it reduces the number of parameters or weights to minimize computational costs. The second step is to prevent the overfitting of the network. In addition, pooling techniques can significantly reduce model training time and computational costs. This paper provides a critical understanding of traditional and modern pooling techniques and highlights the strengths and weaknesses for readers. Moreover, the performance of pooling techniques on different datasets is qualitatively evaluated and reviewed. This study is expected to contribute to a comprehensive understanding of the importance of CNNs and pooling techniques in computer vision challenges.
Oral cancer is considered one of the most common cancer types in several counties. Earlier-stage identification is essential for better prognosis, treatment, and survival. To enhance precision medicine, Internet of Medical Things (IoMT) and deep learning (DL) models can be developed for automated oral cancer classification to improve detection rate and decrease cancer-specific mortality. This article focuses on the design of an optimal Inception-Deep Convolution Neural Network for Oral Potentially Malignant Disorder Detection (OIDCNN-OPMDD) technique in the IoMT environment. The presented OIDCNN-OPMDD technique mainly concentrates on identifying and classifying oral cancer by using an IoMT device-based data collection process. In this study, the feature extraction and classification process are performed using the IDCNN model, which integrates the Inception module with DCNN. To enhance the classification performance of the IDCNN model, the moth flame optimization (MFO) technique can be employed. The experimental results of the OIDCNN-OPMDD technique are investigated, and the results are inspected under specific measures. The experimental outcome pointed out the enhanced performance of the OIDCNN-OPMDD model over other DL models.
Intelligent Transportation Systems (ITS) make use of advanced technologies to optimize interurban and urban traffic, reduce congestion and enhance overall traffic flow. Deep learning (DL) approaches can be widely used for traffic flow monitoring in the ITS. This manuscript introduces the Artificial Hummingbird Optimization Algorithm with Hierarchical Deep Learning for Traffic Management (AHOA-HDLTM) technique in the ITS environment. The purpose of the AHOA-HDLTM technique is to predict traffic flow levels in smart cities, enabling effective traffic management. Primarily, the AHOA-HDLTM model involves data preprocessing and an Improved Salp Swarm Algorithm (ISSA) for feature selection. For the prediction of traffic flow, the Hierarchical Extreme Learning Machine (HELM) model can be used. The HELM extracts complex features and patterns, with an additional Artificial Hummingbird Optimization Algorithm (AHOA)-based hyperparameter selection process to enhance predictive outcomes. The simulation result analysis under different traffic data demonstrates the better performance of the AHOA-HDLTM technique over existing models. The hierarchical structure of the HELM model along with AHOA-based hyperparameter tuning helps to accomplish enhanced prediction performance. The AHOA-HDLTM technique presents a robust solution for traffic management in ITS, showcasing enhanced performance in forecasting traffic patterns and congestion. The AHOA-HDLTM technique can be used in various smart cities and urban regions. Its abilities in real-time traffic flow prediction can be helpful in the design of efficient, sustainable, and resilient transportation networks.
The quantification of corneal endothelial cell (CEC) morphology using manual and semi-automatic software enables an objective assessment of corneal endothelial pathology. However, the procedure is tedious, subjective, and not widely applied in clinical practice. We have developed the CellsDeepNet system to automatically segment and analyse the CEC morphology. The CellsDeepNet system uses Contrast-Limited Adaptive Histogram Equalization (CLAHE) to improve the contrast of the CEC images and reduce the effects of non-uniform image illumination, 2D Double-Density Dual-Tree Complex Wavelet Transform (2DDD-TCWT) to reduce noise, Butterworth Bandpass filter to enhance the CEC edges, and moving average filter to adjust for brightness level. An improved version of U-Net was used to detect the boundaries of the CECs, regardless of the CEC size. CEC morphology was measured as mean cell density (MCD, cell/mm2), mean cell area (MCA, μm2), mean cell perimeter (MCP, μm), polymegathism (coefficient of CEC size variation), and pleomorphism (percentage of hexagonality coefficient). The CellsDeepNet system correlated highly significantly with the manual estimations for MCD (r = 0.94), MCA (r = 0.99), MCP (r = 0.99), polymegathism (r = 0.92), and pleomorphism (r = 0.86), with p < 0.0001 for all the extracted clinical features. The Bland–Altman plots showed excellent agreement. The percentage difference between the manual and automated estimations was superior for the CellsDeepNet system compared to the CEAS system and other state-of-the-art CEC segmentation systems on three large and challenging corneal endothelium image datasets captured using two different ophthalmic devices.
Endoscopic image analysis has played a pivotal function in the diagnosis and management of gastrointestinal (GI) tract diseases. Gastrointestinal endoscopy is a medical procedure where a flexible tube with an endoscope (camera) is inserted into the GI tract to visualize the inner lining of the colon, esophagus, stomach, and small intestine. The videos and images attained during endoscopy provide valuable data for detecting and monitoring a large number of GI diseases. Computer-assisted automated diagnosis technique helps to achieve accurate diagnoses and provide the patient the relevant medical care. Machine learning (ML) and deep learning (DL) methods have been exploited to endoscopic images for classifying diseases and providing diagnostic support. Convolutional Neural Networks (CNN) and other DL algorithms can learn to discriminate between various kinds of GI lesions based on visual properties. This study presents an Endoscopic Image Analysis for Gastrointestinal Tract Disease Diagnosis using an inspired Algorithm with Deep Learning (EIAGTD-NIADL) technique. The EIAGTD-NIADL technique intends to examine the endoscopic images using nature nature-inspired algorithm with a DL model for gastrointestinal tract disease detection and classification. To pre-process the input endoscopic images, the EIAGTD-NIADL technique uses a bilateral filtering (BF) approach. For feature extraction, the EIAGTD-NIADL technique applies an improved ShuffleNet model. To improve the efficacy of the improved ShuffleNet model, the EIAGTD-NIADL technique uses an improved spotted hyena optimizer (ISHO) algorithm. Finally, the classification process is performed by the use of the stacked long short-term memory (SLSTM) method. The experimental outcomes of the EIAGTD-NIADL system can be confirmed on benchmark medical image datasets. The obtained outcomes demonstrate the promising results of the EIAGTD-NIADL approach over other models.
This paper aims to develop a new mobile robot path planning algorithm, called generalized laser simulator (GLS), for navigating autonomously mobile robots in the presence of static and dynamic obstacles. This algorithm enables a mobile robot to identify a feasible path while finding the target and avoiding obstacles while moving in complex regions. An optimal path between the start and target point is found by forming a wave of points in all directions towards the target position considering target minimum and border maximum distance principles. The algorithm will select the minimum path from the candidate points to target while avoiding obstacles. The obstacle borders are regarded as the environment’s borders for static obstacle avoidance. However, once dynamic obstacles appear in front of the GLS waves, the system detects them as new dynamic obstacle borders. Several experiments were carried out to validate the effectiveness and practicality of the GLS algorithm, including path-planning experiments in the presence of obstacles in a complex dynamic environment. The findings indicate that the robot could successfully find the correct path while avoiding obstacles. The proposed method is compared to other popular methods in terms of speed and path length in both real and simulated environments. According to the results, the GLS algorithm outperformed the original laser simulator (LS) method in path and success rate. With application of the all-direction border scan, it outperforms the A-star (A*) and PRM algorithms and provides safer and shorter paths. Furthermore, the path planning approach was validated for local planning in simulation and real-world tests, in which the proposed method produced the best path compared to the original LS algorithm.