The Role of ArtificiaI Intelligence in Brain Tumor Diagnosis: An Evaluation of a Machine Learning Model.

This research study explores of the effectiveness of a machine learning image classification model in the accurate identification of various types of brain tumors. The types of tumors under consideration in this study are gliomas, meningiomas, and pituitary tumors. These are some of the most common types of brain tumors and pose significant challenges in terms of accurate diagnosis and treatment. The machine learning model that is the focus of this study is built on the Google Teachable Machine platform (Alphabet Inc., Mountain View, CA). The Google Teachable Machine is a machine learning image classification platform that is built from Tensorflow, a popular open-source platform for machine learning. The Google Teachable Machine model was specifically evaluated for its ability to differentiate between normal brains and the aforementioned types of tumors in MRI images. MRI images are a common tool in the diagnosis of brain tumors, but the challenge lies in the accurate classification of the tumors. This is where the machine learning model comes into play. The model is trained to recognize patterns in the MRI images that correspond to the different types of tumors. The performance of the machine learning model was assessed using several metrics. These include precision, recall, and F1 score. These metrics were generated from a confusion matrix analysis and performance graphs. A confusion matrix is a table that is often used to describe the performance of a classification model. Precision is a measure of the model's ability to correctly identify positive instances among all instances it identified as positive. Recall, on the other hand, measures the model's ability to correctly identify positive instances among all actual positive instances. The F1 score is a measure that combines precision and recall providing a single metric for model performance. The results of the study were promising. The Google Teachable Machine model demonstrated high performance, with accuracy, precision, recall, and F1 scores ranging between 0.84 and 1.00. This suggests that the model is highly effective in accurately classifying the different types of brain tumors. This study provides insights into the potential of machine learning models in the accurate classification of brain tumors. The findings of this study lay the groundwork for further research in this area and have implications for the diagnosis and treatment of brain tumors. The study also highlights the potential of machine learning in enhancing the field of medical imaging and diagnosis. With the increasing complexity and volume of medical data, machine learning models like the one evaluated in this study could play a crucial role in improving the accuracy and efficiency of diagnoses. Furthermore, the study underscores the importance of continued research and development in this field to further refine these models and overcome any potential limitations or challenges. Overall, the study contributes to the field of medical imaging and machine learning and sets the stage for future research and advancements in this area.

Clinical characteristics and HLA associations of azithromycin-induced liver injury.

BACKGROUND: Azithromycin (AZ) is a widely used antibiotic. The aim of this study was to characterise the clinical features, outcomes, and HLA association in patients with drug-induced liver injury (DILI) due to AZ. METHODS: The clinical characteristics of individuals with definite, highly likely, or probable AZ-DILI enrolled in the US Drug-Induced Liver Injury Network (DILIN) were reviewed. HLA typing was performed using an Illumina MiSeq platform. The allele frequency (AF) of AZ-DILI cases was compared to population controls, other DILI cases, and other antibiotic-associated DILI cases. RESULTS: Thirty cases (4 definite, 14 highly likely, 12 probable) of AZ-DILI were enrolled between 2004 and 2022 with a median age of 46 years, 83% white, and 60% female. Median duration of AZ treatment was 5 days. Latency was 18.5 days. 73% were jaundiced at presentation. The injury pattern was hepatocellular in 60%, cholestatic in 27%, and mixed in 3%. Ten cases (33%) were severe or fatal; 90% of these were hepatocellular. Two patients required liver transplantation. One patient with chronic liver disease died of hepatic failure. Chronic liver injury developed in 17%, of which 80% had hepatocellular injury at onset. HLA-DQA1*03:01 was significantly more common in AZ-DILI versus population controls and amoxicillin-clavulanate DILI cases (AF: 0.29 vs. 0.11, p = 0.001 and 0.002, respectively). CONCLUSION: Azithromycin therapy can lead to rapid onset of severe hepatic morbidity and mortality in adult and paediatric populations. Hepatocellular injury and younger age were associated with worse outcomes. HLA-DQA1*03:01 was significantly more common in AZ cases compared to controls.

Deep learning empowered breast cancer diagnosis: Advancements in detection and classification.

Recent advancements in AI, driven by big data technologies, have reshaped various industries, with a strong focus on data-driven approaches. This has resulted in remarkable progress in fields like computer vision, e-commerce, cybersecurity, and healthcare, primarily fueled by the integration of machine learning and deep learning models. Notably, the intersection of oncology and computer science has given rise to Computer-Aided Diagnosis (CAD) systems, offering vital tools to aid medical professionals in tumor detection, classification, recurrence tracking, and prognosis prediction. Breast cancer, a significant global health concern, is particularly prevalent in Asia due to diverse factors like lifestyle, genetics, environmental exposures, and healthcare accessibility. Early detection through mammography screening is critical, but the accuracy of mammograms can vary due to factors like breast composition and tumor characteristics, leading to potential misdiagnoses. To address this, an innovative CAD system leveraging deep learning and computer vision techniques was introduced. This system enhances breast cancer diagnosis by independently identifying and categorizing breast lesions, segmenting mass lesions, and classifying them based on pathology. Thorough validation using the Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM) demonstrated the CAD system's exceptional performance, with a 99% success rate in detecting and classifying breast masses. While the accuracy of detection is 98.5%, when segmenting breast masses into separate groups for examination, the method's performance was approximately 95.39%. Upon completing all the analysis, the system's classification phase yielded an overall accuracy of 99.16% for classification. The potential for this integrated framework to outperform current deep learning techniques is proposed, despite potential challenges related to the high number of trainable parameters. Ultimately, this recommended framework offers valuable support to researchers and physicians in breast cancer diagnosis by harnessing cutting-edge AI and image processing technologies, extending recent advances in deep learning to the medical domain.

Safety of robotic cholecystectomy as index training procedure: the UK experience.

AIMS: To evaluate the safety profile of robotic cholecystectomy performed within the United Kingdom (UK) Robotic Hepatopancreatobiliary (HPB) training programme. METHODS: A retrospective evaluation of prospectively collected data from eleven centres participating in the UK Robotic HPB training programme was conducted. All adult patients undergoing robotic cholecystectomy for symptomatic gallstone disease or gallbladder polyp were considered. Bile duct injury, conversion to open procedure, conversion to subtotal cholecystectomy, length of hospital stay, 30-day re-admission, and post-operative complications were the evaluated outcome parameters. RESULTS: A total of 600 patients were included. The median age was 53 (IQR 65-41) years and the majority (72.7%; 436/600) were female. The main indications for robotic cholecystectomy were biliary colic (55.5%, 333/600), cholecystitis (18.8%, 113/600), gallbladder polyps (7.7%, 46/600), and pancreatitis (6.2%, 37/600). The median length of stay was 0 (IQR 0-1) days. Of the included patients, 88.5% (531/600) were discharged on the day of procedure with 30-day re-admission rate of 5.5% (33/600). There were no bile duct injuries and the rate of conversion to open was 0.8% (5/600) with subtotal cholecystectomy rate of 0.8% (5/600). CONCLUSION: The current study confirms that robotic cholecystectomy can be safely implemented to routine practice with a low risk of bile duct injury, low bile leak rate, low conversion to open surgery, and low need for subtotal cholecystectomy.

Characterisation of an outbreak of acute liver injury after ingestion of plant-based food supplement.

BACKGROUND: In April 2022, French Lentil and Leek Crumble (FLLC), a new frozen food preparation manufactured by Daily Harvest™ (containing Tara flour) was offered as a natural high-protein meal product. Soon thereafter, widespread anecdotal reports of acute gastrointestinal symptoms with liver injury were reported, leading to its voluntary withdrawal in June 2022, after shipment of 28,000 preparations. AIMS: To summarise the clinical and laboratory features of 17 patients with FLLC associated liver injury from the Drug Induced Liver Injury Network (DILIN). METHODS: Patients with FLLC-associated liver injury were enrolled into a prospective protocol and followed for 6 months. Cases were adjudicated by expert opinion causality assessment with summary statistics for data analysis. RESULTS: Enrolled subjects had a mean age of 41 years, 82% were female with mean BMI of 24 kg/m2. All were Caucasian without underlying liver disease. In most cases, abdominal pain and nausea arose within hours of FLLC ingestion. Mean days from ingestion to identification of liver injury was 3.1 days (±2.8). On enrolment, 53% had jaundice, 47% nausea, 24% fever, 59% abdominal pain, 41% itching and 12% rash. The mean initial serum ALT was 475 U/L (±302), AST 315 U/L (±315), alkaline phosphatase 190 U/L (±76), with a total bilirubin of 2.6 mg/dL (±2). In this study, 63% presented with a hepatocellular pattern of liver injury, 6% cholestatic and 31% mixed as determined by the R value. In addition, 24% of patients were hospitalised, and there were no fatalities or liver transplants. Liver biopsy in one subject revealed acute hepatitis with mild ductular reaction, mild lymphocytic and eosinophilic portal inflammation, mild lobular inflammation, preserved bile ducts and absence of interface hepatitis, steatosis, granulomatous reaction or cholestasis. Phylogenetic analysis confirmed the presence of Tara spinosa, the source of Tara flour. CONCLUSIONS: Natural food products are increasingly ubiquitous and may unexpectedly cause significant illness. All clinicians should inquire whether patients are consuming natural food products or herbal supplements and consider them as a potential cause of liver injury.

A transformer-based approach empowered by a self-attention technique for semantic segmentation in remote sensing.

Semantic segmentation of Remote Sensing (RS) images involves the classification of each pixel in a satellite image into distinct and non-overlapping regions or segments. This task is crucial in various domains, including land cover classification, autonomous driving, and scene understanding. While deep learning has shown promising results, there is limited research that specifically addresses the challenge of processing fine details in RS images while also considering the high computational demands. To tackle this issue, we propose a novel approach that combines convolutional and transformer architectures. Our design incorporates convolutional layers with a low receptive field to generate fine-grained feature maps for small objects in very high-resolution images. On the other hand, transformer blocks are utilized to capture contextual information from the input. By leveraging convolution and self-attention in this manner, we reduce the need for extensive downsampling and enable the network to work with full-resolution features, which is particularly beneficial for handling small objects. Additionally, our approach eliminates the requirement for vast datasets, which is often necessary for purely transformer-based networks. In our experimental results, we demonstrate the effectiveness of our method in generating local and contextual features using convolutional and transformer layers, respectively. Our approach achieves a mean dice score of 80.41%, outperforming other well-known techniques such as UNet, Fully-Connected Network (FCN), Pyramid Scene Parsing Network (PSP Net), and the recent Convolutional vision Transformer (CvT) model, which achieved mean dice scores of 78.57%, 74.57%, 73.45%, and 62.97% respectively, under the same training conditions and using the same training dataset.

Robotic Versus Laparoscopic Liver Resection in Various Settings: An International Multicenter Propensity Score Matched Study of 10.075 Patients.

OBJECTIVE: To compare the perioperative outcomes of robotic liver surgery (RLS) and laparoscopic liver surgery (LLS) in various settings. BACKGROUND: Clear advantages of RLS over LLS have rarely been demonstrated, and the associated costs of robotic surgery are generally higher than those of laparoscopic surgery. Therefore, the exact role of the robotic approach in minimally invasive liver surgery remains to be defined. METHODS: In this international retrospective cohort study, the outcomes of patients who underwent RLS and LLS for all indications between 2009 and 2021 in 34 hepatobiliary referral centers were compared. Subgroup analyses were performed to compare both approaches across several types of procedures: (1) minor resections in the anterolateral (2, 3, 4b, 5, and 6) or (2) posterosuperior segments (1, 4a, 7, 8), and (3) major resections (≥3 contiguous segments). Propensity score matching was used to mitigate the influence of selection bias. The primary outcome was textbook outcome in liver surgery (TOLS), previously defined as the absence of intraoperative incidents ≥grade 2, postoperative bile leak ≥grade B, severe morbidity, readmission, and 90-day or in-hospital mortality with the presence of an R0 resection margin in case of malignancy. The absence of a prolonged length of stay was added to define TOLS+. RESULTS: Among the 10.075 included patients, 1.507 underwent RLS and 8.568 LLS. After propensity score matching, both groups constituted 1.505 patients. RLS was associated with higher rates of TOLS (78.3% vs 71.8%, P < 0.001) and TOLS+ (55% vs 50.4%, P = 0.026), less Pringle usage (39.1% vs 47.1%, P < 0.001), blood loss (100 vs 200 milliliters, P < 0.001), transfusions (4.9% vs 7.9%, P = 0.003), conversions (2.7% vs 8.8%, P < 0.001), overall morbidity (19.3% vs 25.7%, P < 0.001), and microscopically irradical resection margins (10.1% vs. 13.8%, P = 0.015), and shorter operative times (190 vs 210 minutes, P = 0.015). In the subgroups, RLS tended to have higher TOLS rates, compared with LLS, for minor resections in the posterosuperior segments (n = 431 per group, 75.9% vs 71.2%, P = 0.184) and major resections (n = 321 per group, 72.9% vs 67.5%, P = 0.086), although these differences did not reach statistical significance. CONCLUSIONS: While both produce excellent outcomes, RLS might facilitate slightly higher TOLS rates than LLS.

Recent advances in the diagnosis of drug-induced liver injury.

Drug-induced liver injury (DILI) is a major problem in the United States, commonly leading to hospital admission. Diagnosing DILI is difficult as it is a diagnosis of exclusion requiring a temporal relationship between drug exposure and liver injury and a thorough work up for other causes. In addition, DILI has a very variable clinical and histologic presentation that can mimic many different etiologies of liver disease. Objective scoring systems can assess the probability that a drug caused the liver injury but liver biopsy findings are not part of the criteria used in these systems. This review will address some of the recent updates to the scoring systems and the role of liver biopsy in the diagnosis of DILI.

Meta-analysis and trial sequential analysis of pancreatic stump closure using a hand-sewn or stapler technique in distal pancreatectomy.

This study aimed to compare outcomes of hand-sewn and stapler closure techniques of pancreatic stump in patients undergoing distal pancreatectomy (DP). Impact of stapler closure reinforcement using mesh on outcomes was also evaluated. Literature search was carried out using multiple data sources to identify studies that compared hand-sewn and stapler closure techniques in management of pancreatic stump following DP. Odds ratio (OR) was determined for clinically relevant postoperative pancreatic fistula (POPF) via random-effects modelling. Subsequently, trial sequential analysis was performed. Thirty-two studies with a total of 4,022 patients undergoing DP with hand-sewn (n = 1,184) or stapler (n = 2,838) closure technique of pancreatic stump were analyzed. Hand-sewn closure significantly increased the risk of clinically relevant POPF compared to stapler closure (OR: 1.56, p = 0.02). When stapler closure was considered, staple line reinforcement significantly reduced formation of such POPF (OR: 0.54, p = 0.002). When only randomized controlled trials were considered, there was no significant difference in clinically relevant POPF between hand-sewn and stapler closure techniques (OR: 1.20, p = 0.64) or between reinforced and standard stapler closure techniques (OR: 0.50, p = 0.08). When observational studies were considered, hand-sewn closure was associated with a significantly higher rate of clinically relevant POPF compared to stapler closure (OR: 1.59, p = 0.03). Moreover, when stapler closure was considered, staple line reinforcement significantly reduced formation of such POPF (OR: 0.55, p = 0.02). Trial sequential analysis detected risk of type 2 error. In conclusion, reinforced stapler closure in DP may reduce risk of clinically relevant POPF compared to hand-sewn closure or stapler closure without reinforcement. Future randomized research is needed to provide stronger evidence.

Perspectives on robotic HPB training in the UK: a survey analysis.

BACKGROUND: We Published a step-up approach for robotic training in hepato-pancreato-biliary (HPB) surgery has been previously. The approach was mostly based on personal experience and communications between experts and needed appraisal and validation by the HPB surgical community. At the Great Britain and Ireland HPB Association (GBIHPBA) robotic HPB meeting held in Coventry, UK in October 2022, the authors sought consensus from the live audience, with an open forum for answering key questions. The aim of this exercise was to appraise the step-up approach, and in turn, lay the foundation for a more substantial UK robotic HPB surgical curriculum. METHODS: The study was conducted using VEVOX online polling platform at the October 2022 GBIHPBA robotic HPB meeting in Coventry, UK. The questionnaire was designed based on a literature search and was externally validated. The data were collated and analysed to assess patterns of response. RESULTS: A median (IQR) of 104 (96-117) responses were generated for each question. 93 consultants and 61 trainees were present Over 90% were in favour of a standardised training pathway. 93.6% were in favour of the proposed step-up approach, with a significant number (67.3%; p < 0.001) considering three levels of case complexity. CONCLUSION: The survey shows a favourable outlook on adopting step-up training in robotic HPB surgery. Ongoing monitoring of progress, clinical outcomes, and collaboration among surgeons and units will bolster this evidence, potentially leading to an official UK robotic HPB curriculum.

Amiodarone and Dronedarone Causes Liver Injury with Distinctly Different Clinical Presentations.

OBJECTIVE: To describe hepatotoxicity due to amiodarone and dronedarone from the DILIN and the US FDA's surveillance database. METHODS: Hepatotoxicity due to amiodarone and dronedarone enrolled in the U.S. Drug Induced Liver Injury Network (DILIN) from 2004 to 2020 are described. Dronedarone hepatotoxicity cases associated with liver biopsy results were obtained from the FDA Adverse Event Reporting System (FAERS) from 2009 to 2020. RESULTS: Among DILIN's 10 amiodarone and 3 dronedarone DILIN cases, the latency for amiodarone was longer than with dronedarone (388 vs 119 days, p = 0.50) and the median ALT at DILI onset was significantly lower with amiodarone (118 vs 1191 U/L, p = 0.05). Liver biopsies in five amiodarone cases showed fibrosis, steatosis, and numerous Mallory-Denk bodies. Five patients died although only one from liver failure. One patient with dronedarone induced liver injury died of a non-liver related cause. Nine additional cases of DILI due to dronedarone requiring hospitalization were identified in the FAERS database. Three patients developed liver injury within a month of starting the medication. Two developed acute liver failure and underwent urgent liver transplant, one was evaluated for liver transplant but then recovered spontaneously, while one patient with cirrhosis died of liver related causes. CONCLUSION: Amiodarone hepatotoxicity resembles that seen in alcohol related liver injury, with fatty infiltration and inflammation. Dronedarone is less predictable, typically without fat and with a shorter latency of use before presentation. These differences may be explained, in part, by the differing pharmacokinetics of the two drugs leading to different mechanisms of hepatotoxicity.

Comparative study of operative expenses: robotic vs. laparoscopic vs. open liver resections at a university hospital in the UK.

Robotic liver resections (RLR) are increasingly being performed and has previously been considered more costly. The aim is to explore the cost of RLR compared with laparoscopic and open liver resection in a single National Health Service (NHS) hospital. A retrospective review of patients who underwent RLR, LLR, and OLR from April 2014 to December 2022 was conducted. The primary outcomes were the cost of consumables and median income, and the secondary outcomes were the overall length of stay and mortality at 90 days. Overall, 332 patients underwent liver resections. There were 204 males (61.4%) and 128 females (38.6%), with a median age of 62 years (IQR: 51-77 years). Of these, 60 patients (18.1%) underwent RLR, 21 patients (6.3%) underwent LLR, and 251 patients (75.6%) underwent OLR. Median consumables cost per case was £3863 (IQR: £3458-£5061) for RLR, £4326 (IQR: £4273-£4473) for LLR, and £4,084 (IQR: £3799-£5549) for the OLR cohort (p = 0.140). Median income per case was £7999 (IQR: £4509-£10,777) for RLR, £7497 (IQR: £2407-£14,576) for LLR, and £7493 (IQR: £2542-£14,121) for OLR. The median length of stay (LOS) for RLR was 3 days (IQR: 2-4.7 days) compared to 5 days for LLR (IQR: 4.5-7 days) and 6 days for OLR (IQR: 5-8 days, p < 0.001). Within the NHS, RLR has consumable costs comparable to OLR and LLR. It is also linked with a shorter LOS and generates similar income for patients undergoing OLR and LLR.

The Brescia Internationally Validated European Guidelines on Minimally Invasive Pancreatic Surgery (EGUMIPS).

OBJECTIVE: To develop and update evidence-based and consensus-based guidelines on laparoscopic and robotic pancreatic surgery. SUMMARY BACKGROUND DATA: Minimally invasive pancreatic surgery (MIPS), including laparoscopic and robotic surgery, is complex and technically demanding. Minimizing the risk for patients requires stringent, evidence-based guidelines. Since the International Miami Guidelines on MIPS in 2019, new developments and key publications have been reported, necessitating an update. METHODS: Evidence-based guidelines on 22 topics in 8 domains were proposed: terminology, indications, patients, procedures, surgical techniques and instrumentation, assessment tools, implementation and training, and artificial intelligence. The Brescia Internationally Validated European Guidelines on Minimally Invasive Pancreatic Surgery (EGUMIPS, September 2022) used the Scottish Intercollegiate Guidelines Network (SIGN) methodology to assess the evidence and develop guideline recommendations, the Delphi method to establish consensus on the recommendations among the Expert Committee, and the AGREE II-GRS tool for guideline quality assessment and external validation by a Validation Committee. RESULTS: Overall, 27 European experts, 6 international experts, 22 international Validation Committee members, 11 Jury Committee members, 18 Research Committee members, and 121 registered attendees of the 2-day meeting were involved in the development and validation of the guidelines. In total, 98 recommendations were developed, including 33 on laparoscopic, 34 on robotic, and 31 on general MIPS, covering 22 topics in 8 domains. Out of 98 recommendations, 97 reached at least 80% consensus among the experts and congress attendees, and all recommendations were externally validated by the Validation Committee. CONCLUSIONS: The EGUMIPS evidence-based guidelines on laparoscopic and robotic MIPS can be applied in current clinical practice to provide guidance to patients, surgeons, policy-makers, and medical societies.

Bouveret's syndrome: An old diagnosis. A modern multimodality approach (endoscopic and robotic surgical) of gastric outlet obstruction: Report of two cases.

INTRODUCTION AND IMPORTANCE: Bouveret's syndrome is an uncommon condition characterized by the impaction of a gallstone in the pylorus or duodenum via a cholecysto-enteric fistula causing gastric outlet obstruction. We report two unusual cases of Bouveret's syndrome causing gastric outlet obstruction in two elderly patients. CASE PRESENTATION: Two elderly female patients presented to the surgical assessment unit with features of gastric outlet obstruction. In both cases, an urgent computed tomography (CT) of the abdomen showed pneumobilia, gastric distension, and gallstones impaction at the duodenal bulb. In Patient 1, endoscopic removal of the impacted gallstones was done successfully. She was discharged three days following an uneventful recovery. In Patient 2, an endoscopic removal of a single large gallstone was attempted, which was unsuccessful. She underwent robotic gastrotomy with extraction of the large gallstone with primary repair. She was discharged on 8th postoperative day. CLINICAL DISCUSSION: Treatment options for Bouveret's syndrome include endoscopic management and surgery. The selection of treatment options depends upon factors like the degree of obstruction, the impaction site, number, type or size of gallstones, patient co-morbidities and clinical parameters at presentation, as well as expertise available, both endoscopic and surgical. CONCLUSIONS: Bouveret's syndrome is one of the rare complications of gallstone. Endoscopic management can be effective at removing the impacted gallstones, which is particularly helpful for those elderly patients who have multiple medical co-morbidities, as in our first patient. Surgical management like minimal invasive surgery (robotic) can be beneficial in failed endoscopic attempt of removal of stone like in the second patient.

Automatic neonatal sleep stage classification: A comparative study.

Sleep is an essential feature of living beings. For neonates, it is vital for their mental and physical development. Sleep stage cycling is an important parameter to assess neonatal brain and physical development. Therefore, it is crucial to administer newborn's sleep in the neonatal intensive care unit (NICU). Currently, Polysomnography (PSG) is used as a gold standard method for classifying neonatal sleep patterns, but it is expensive and requires a lot of human involvement. Over the last two decades, multiple researchers are working on automatic sleep stage classification algorithms using electroencephalography (EEG), electrocardiography (ECG), and video. In this study, we present a comprehensive review of existing algorithms for neonatal sleep, their limitations and future recommendations. Additionally, a brief comparison of the extracted features, classification algorithms and evaluation parameters is reported in the proposed study.

Comparative Analysis of Machine Learning Models for Image Detection of Colonic Polyps vs. Resected Polyps.

(1) Background: Colon polyps are common protrusions in the colon's lumen, with potential risks of developing colorectal cancer. Early detection and intervention of these polyps are vital for reducing colorectal cancer incidence and mortality rates. This research aims to evaluate and compare the performance of three machine learning image classification models' performance in detecting and classifying colon polyps. (2) Methods: The performance of three machine learning image classification models, Google Teachable Machine (GTM), Roboflow3 (RF3), and You Only Look Once version 8 (YOLOv8n), in the detection and classification of colon polyps was evaluated using the testing split for each model. The external validity of the test was analyzed using 90 images that were not used to test, train, or validate the model. The study used a dataset of colonoscopy images of normal colon, polyps, and resected polyps. The study assessed the models' ability to correctly classify the images into their respective classes using precision, recall, and F1 score generated from confusion matrix analysis and performance graphs. (3) Results: All three models successfully distinguished between normal colon, polyps, and resected polyps in colonoscopy images. GTM achieved the highest accuracies: 0.99, with consistent precision, recall, and F1 scores of 1.00 for the 'normal' class, 0.97-1.00 for 'polyps', and 0.97-1.00 for 'resected polyps'. While GTM exclusively classified images into these three categories, both YOLOv8n and RF3 were able to detect and specify the location of normal colonic tissue, polyps, and resected polyps, with YOLOv8n and RF3 achieving overall accuracies of 0.84 and 0.87, respectively. (4) Conclusions: Machine learning, particularly models like GTM, shows promising results in ensuring comprehensive detection of polyps during colonoscopies.

AI-Enabled Traffic Control Prioritization in Software-Defined IoT Networks for Smart Agriculture.

Smart agricultural systems have received a great deal of interest in recent years because of their potential for improving the efficiency and productivity of farming practices. These systems gather and analyze environmental data such as temperature, soil moisture, humidity, etc., using sensor networks and Internet of Things (IoT) devices. This information can then be utilized to improve crop growth, identify plant illnesses, and minimize water usage. However, dealing with data complexity and dynamism can be difficult when using traditional processing methods. As a solution to this, we offer a novel framework that combines Machine Learning (ML) with a Reinforcement Learning (RL) algorithm to optimize traffic routing inside Software-Defined Networks (SDN) through traffic classifications. ML models such as Logistic Regression (LR), Random Forest (RF), k-nearest Neighbours (KNN), Support Vector Machines (SVM), Naive Bayes (NB), and Decision Trees (DT) are used to categorize data traffic into emergency, normal, and on-demand. The basic version of RL, i.e., the Q-learning (QL) algorithm, is utilized alongside the SDN paradigm to optimize routing based on traffic classes. It is worth mentioning that RF and DT outperform the other ML models in terms of accuracy. Our results illustrate the importance of the suggested technique in optimizing traffic routing in SDN environments. Integrating ML-based data classification with the QL method improves resource allocation, reduces latency, and improves the delivery of emergency traffic. The versatility of SDN facilitates the adaption of routing algorithms depending on real-time changes in network circumstances and traffic characteristics.

A hybrid deep learning-based intrusion detection system for IoT networks.

The Internet of Things (IoT) is a rapidly evolving technology with a wide range of potential applications, but the security of IoT networks remains a major concern. The existing system needs improvement in detecting intrusions in IoT networks. Several researchers have focused on intrusion detection systems (IDS) that address only one layer of the three-layered IoT architecture, which limits their effectiveness in detecting attacks across the entire network. To address these limitations, this paper proposes an intelligent IDS for IoT networks based on deep learning algorithms. The proposed model consists of a recurrent neural network and gated recurrent units (RNN-GRU), which can classify attacks across the physical, network, and application layers. The proposed model is trained and tested using the ToN-IoT dataset, specifically collected for a three-layered IoT system, and includes new types of attacks compared to other publicly available datasets. The performance analysis of the proposed model was carried out by a number of evaluation metrics such as accuracy, precision, recall, and F1-measure. Two optimization techniques, Adam and Adamax, were applied in the evaluation process of the model, and the Adam performance was found to be optimal. Moreover, the proposed model was compared with various advanced deep learning (DL) and traditional machine learning (ML) techniques. The results show that the proposed system achieves an accuracy of 99% for network flow datasets and 98% for application layer datasets, demonstrating its superiority over previous IDS models.

Multi-step attack detection in industrial networks using a hybrid deep learning architecture.

In recent years, the industrial network has seen a number of high-impact attacks. To counter these threats, several security systems have been implemented to detect attacks on industrial networks. However, these systems solely address issues once they have already transpired and do not proactively prevent them from occurring in the first place. The identification of malicious attacks is crucial for industrial networks, as these attacks can lead to system malfunctions, network disruptions, data corruption, and the theft of sensitive information. To ensure the effectiveness of detection in industrial networks, which necessitate continuous operation and undergo changes over time, intrusion detection algorithms should possess the capability to automatically adapt to these changes. Several researchers have focused on the automatic detection of these attacks, in which deep learning (DL) and machine learning algorithms play a prominent role. This study proposes a hybrid model that combines two DL algorithms, namely convolutional neural networks (CNN) and deep belief networks (DBN), for intrusion detection in industrial networks. To evaluate the effectiveness of the proposed model, we utilized the Multi-Step Cyber Attack (MSCAD) dataset and employed various evaluation metrics.