Test Accuracy Research Articles

Comparison of the accuracy of OES and EDX tests on nickel dissolving in brass casting

Comparison of the accuracy of OES and EDX tests on nickel dissolving in brass casting

Mitigating Gradient-Based Data Poisoning Attacks on Machine Learning Models: A Statistical Detection Method

Objectives: This research paper aims to develop a novel method for identifying gradient-based data poisoning attacks on industrial applications like autonomous vehicles and intelligent healthcare systems relying on machine learning and deep learning techniques. These algorithms performs well only if they are trained on good quality dataset. However, the ML models are prone to data poisoning attacks, targeting the training dataset, manipulate its input samples such that the machine learning algorithm gets confused and produces wrong predictions. The current detection techniques are effective to detect known attacks and lack generalized detection to unknown attacks. To address this issue, this paper aims to integrate security elements within the machine learning framework, guaranteeing effective identification and mitigation of known and unknown threats and achieve generalized detection. Methods: ML Filter, a unique attack detection approach integrates ML-Filter Detection Algorithm and the Statistical Perturbation Bounds Identification Algorithm to determine the given dataset is poisoned or not. DBSCAN algorithm is used to divide the dataset into several smaller subsets and perform algorithmic analysis for detection. The performance of the proposed method is evaluated in terms of True positive rate and significance test accuracy. Findings: The probability distribution differences between original and poisoned datasets vary with change in perturbation size rather than the datasets and ML models use for application. This finding lead to determine the perturbation bounds using statistical pairwise distance metrics and corresponding significance tests computed on the results. ML Filter demonstrates a high detection rate of 99.63% for known attacks and achieves a generalized detection accuracy of 98% for unknown attacks. Novelty: A secured ML architecture and a unique statistical detection approach ML-Filter, effectively detect data poisoning attacks, demonstrating significant advancements in detecting both known and unknown threats in industrial applications utilizing machine learning and deep learning algorithms. Keywords: Privacy and security, Adversarial machine learning, Secured ML Architecture, ML-Filter, Statistical Perturbation Bounds Identification Algorithm

Experimental Study of Interference in Calcitonin Testing: A Case Report and Literature Review.

Calcitonin (CT) is a sensitive serum marker of medullary thyroid carcinoma usually detected via immunoassays; however, its levels are easily disturbed by several endogenous factors. The study aimed to discuss a case of suspected interference resulting in aberrant CT values and review previous reports of CT interference. A female patient visited our clinic with a physical ultrasound examination showing a slightly enlarged thyroid gland with small nodules. She had elevated CT levels, inconsistent with the clinical presentation and other findings. We evaluated the results by retesting using the same platform, platform validation, multiplex dilution, Polyethylene Glycol (PEG) precipitation, heterophilic blocking tubes, and RET gene analysis. Retesting CT using the same platform confirmed the high value obtained. However, serial dilution of the sample produced nonlinear results, suggesting some interference. While PEG precipitation did not significantly reduce the CT level, incubating the sample in HBTs normalized the CT value, indicating interference from heterophilic antibodies. Gene sequencing revealed no RET mutations. In cases where elevated CT levels are inconsistent with clinical presentations and other findings, the laboratory technicians should communicate with clinicians, analyze the reasons for the inconsistent results, and use different methods to verify the results. Accurate testing provides realistic and reliable data for doctors and patients and helps to avoid unnecessary procedures.

DROPLET MOTION ON HYDROPHOBIC AND HYDROPHILIC SURFACES USING LATTICE BOLTZMANN METHOD WITH BLOOD TEST MICROPAD APPLICATION

Laboratory sciences and disease diagnostics need accelerated medical tests, ease of testing for patients, and lower diagnostic costs. In this respect, one of the most innovative methods is paper-based microfluidic analysis tools, called micro-PAD. Since the chemical-based micro-PADs need sufficient blood to show an accurate blood test, the shape of the micro-PADs, the blood quantity, and the droplet flow to reach the end of the microchannels are important. Therefore, this research showed the liquid droplet motion on a heterogeneous (hydrophobic and hydrophilic) surface using the lattice Boltzmann method, and two star-shaped micro-PADs were modeled to investigate the effect of the droplet size and the droplet deflection from the center of the micro-PAD on liquid flow in microchannels. Based on the obtained results, fluid behavior was similar in two patterns (ten-pointed pattern and four-pointed pattern). This study compared the effects of parallel, convergent, and divergent channels in a four-pointed pattern on fluid velocity. The off-center error of the droplet was considered at the initial moment. The obtained results confirmed the direct relationship between the increased deviation of the center and the lack of uniformity in the fluid penetration. The present simulation results were validated against the laboratory data.

Sliding motility of Bacillus cereus mediates vancomycin pseudo-resistance during antimicrobial susceptibility testing.

The glycopeptide vancomycin is the antimicrobial agent-of-choice for the treatment of severe non-gastrointestinal infections with members of Bacillus cereus sensu lato (s.l.). Recently, sporadic detection of vancomycin-resistant phenotypes emerged, mostly for agar diffusion testing such as the disc diffusion method or gradient test (e.g. Etest®) method. In this work, we were able to disprove a preliminarily assumed high resistance to vancomycin in an isolate of B. cereus s.l. using broth microdilution and agar dilution. Microscopic imaging during vancomycin susceptibility testing showed spreading towards the inhibition zone, which strongly suggested sliding motility. Furthermore, transcriptomic analysis using RNA-Seq on the nanopore platform revealed several key genes of biofilm formation (e.g. calY, tasA, krsEABC) to be up-regulated in pseudo-resistant cells, substantiating that bacterial sliding is responsible for the observed mobility. Down-regulation of virulence (e.g. hblABCD, nheABC, plcR) and flagellar genes compared with swarming cells also confirmed the non-swarming phenotype of the pseudo-resistant isolate. The results highlight an insufficiency of agar diffusion testing for vancomycin susceptibility in the B. cereus group, and reference methods like broth microdilution are strongly recommended. As currently no guideline mentions interfering phenotypes in antimicrobial susceptibility testing of B. cereus s.l., this knowledge is essential to obtain reliable results on vancomycin susceptibility. In addition, this is the first report of sliding motility undermining accurate antimicrobial susceptibility testing in B. cereus s.l. and may serve as a basis for future studies on bacterial motility in susceptibility testing and its potential impact on treatment efficacy.

Summary of Research on Highway Bridge Vehicle Force Identification

Vehicle force identification is one of the core technical problems to be solved urgently in the management of transportation infrastructure, and it has also been a research hotspot in recent years. To promote the application of vehicle force identification technology on bridges and explore its development direction, the development status of indirect vehicle force identification methods based on bridge response is reviewed during this study. The basic theories of two major methods, including bridge weigh-in-motion (BWIM) and moving force identification (MFI), are described in detail in this study, and then, the key technical principles of bridge force identification are revealed. Secondly, the development status of BWIM in recent years is reviewed from three aspects, including test accuracy, applicability and test efficiency. Combined with a variety of theories, the current status of MFI is analyzed from the establishment of the solution to the equation. Finally, the development direction of an artificial neural network and machine vision technology are prospected in this study. The BP neural network has good self-learning ability and self-adaptive ability, but the algorithm needs to be improved. The identification method based on machine vision represents the current development direction in vehicle force identification, with great potential.

Predictive machine learning model for mechanical dilatation in transvenous lead extraction procedures

Abstract Background Transvenous lead extraction (TLE) remains a procedure that requires a high level of expertise, with a doubled risk of death and clinical failure when performed in low-volume centers compared to high-volume ones (1). Purpose The aim of this study was to create a machine learning (ML)-based risk stratification system for predicting the need for mechanical dilatation in patients undergoing TLE due to infection. Methods We designed a ML-based risk stratification system trained with data from our registry to predict the need for mechanical dilatation in patients undergoing TLE for infection. An extensive evaluation of 5 different ML models (k-nearest neighbors, support vector machine, decision tree, and decision tree ensembles, such as random forest and gradient boosting machine) was conducted to identify a classifier with the highest potential to correctly predict previously unseen patients. Results Data to train the model was extracted from our 25-year registry of patients undergoing TLE (June 1998 – March 2023), for a total of 491 patients (77.8% male; age 69.7 ± 12.8 years) and 938 leads (ICD 21.2%; pacing 78.8%; indwelling time 61 ± 60 months) removed with success in 100% of cases. Each patient was represented by a set of 21 attributes (14 clinical, 7 device-related). Manual traction (MT) was used in 27.5% of cases, and mechanical dilatation (MD) was employed in the remaining 72.5% of cases. 5-fold nested cross validation was used to estimate performances: in turn, 393 patients were used for training and model selection, and 98 patients were used for independent testing. According to the evaluation, Gradient Boosting Machine performed best, achieving test accuracy of 89% (+/- 2% std. dev.), test sensitivity of 95% (+/- 3% std. dev.), test specificity of 73% (+/- 8% std. dev.), test AUROC of 92% (+/- 1% std. dev.). A further interpretability analysis on the best performing decision tree was conducted, showing remarkable adherence between the internal decisions taken by the model to make predictions and the current clinical practice for TLE. Conclusion ML models have enhanced the prediction of MD requirements in TLE procedures. This data may support the referral of patients to specialized and high-volume centers, thereby improving preoperative risk assessment.Feature importanceROC Curves

Accelerated construction of stress relief music datasets using CNN and the Mel-scaled spectrogram.

Listening to music is a crucial tool for relieving stress and promoting relaxation. However, the limited options available for stress-relief music do not cater to individual preferences, compromising its effectiveness. Traditional methods of curating stress-relief music rely heavily on measuring biological responses, which is time-consuming, expensive, and requires specialized measurement devices. In this paper, a deep learning approach to solve this problem is introduced that explicitly uses convolutional neural networks and provides a more efficient and economical method for generating large datasets of stress-relief music. These datasets are composed of Mel-scaled spectrograms that include essential sound elements (such as frequency, amplitude, and waveform) that can be directly extracted from the music. The trained model demonstrated a test accuracy of 98.7%, and a clinical study indicated that the model-selected music was as effective as researcher-verified music in terms of stress-relieving capacity. This paper underlines the transformative potential of deep learning in addressing the challenge of limited music options for stress relief. More importantly, the proposed method has profound implications for music therapy because it enables a more personalized approach to stress-relief music selection, offering the potential for enhanced emotional well-being.

#2124 Accuracy and interferences in the T50 test: assessing the reliability of a novel diagnostic tool

Abstract Background and Aims A novel blood-based in vitro diagnostic test determines the calcification propensity in patients, e.g. chronic kidney disease (CKD) patients. This assay is known as the ‘T50 test’ and is based on the one-half maximal transition time (T50; minutes) of calciprotein particle 1 (CPP1) to calciprotein particle 2 (CPP2) in serum (Fig. 1). Numerous studies have associated a lower T50 with cardiovascular and all-cause mortality in chronic kidney disease (CKD) patients at stages 3, 4 and 5, kidney transplant recipients, and in the general population. To date, the T50 test has been performed at a single site (Calciscon AG, Biel, Switzerland). The aim of this study was to determine the accuracy of the T50 test. Additionally, the effect of haemolysis, hyperbilirubinemia and lipidemia interferences on the T50 was studied and compared to the manufacturer's claims. Method The T50 was measured using T50 test kits and according to the protocol of Calciscon AG using an automatic pipetting machine (EVOware Freedom, Tecan) and a nephelometer (NEPHELOstar, BMG). Patient serum samples were spiked with haemoglobin, bilirubin, or lipids (triglyceride) for the interference analysis. In total, three dilution series with interferences and four additional non-spiked patient serum samples were included (n = 32). The T50 of all samples were measured at two laboratory sites: Amsterdam UMC (Amsterdam, The Netherlands) and Calciscon AG. Data were statistically analysed for accuracy using the Passing Bablok regression and the Bland Altman plot. Interference was established with deviations >10% in T50 compared to the control sample. Results The measurements of all samples at both sites were similar with a mean bias of 0.24% (±4.84%) and a great agreement was found (slope = 1.012; 95% CI [0.931 to 1.107]) (Table 1). Interference was observed at higher levels of haemoglobin and triglyceride with concentrations of at least 0.1 g/dL and 1124.8 mg/dL, respectively. The triglyceride interference is in agreement with the manufacturer's claim (>1000 mg/dL) in contrary to haemoglobin (>0.3 g/dL). No interference was detected for bilirubin with concentrations up to 59.8 mg/dL. Conclusion The accuracy of the T50 test was high with nearly identical T50 results between two laboratory sites. Furthermore, no interference was observed within usual clinical ranges of haemoglobin, bilirubin, and lipid levels. Interference was present at extremely high values only.

Enhancing trash classification in smart cities using federated deep learning

Efficient Waste management plays a crucial role to ensure clean and green environment in the smart cities. This study investigates the critical role of efficient trash classification in achieving sustainable solid waste management within smart city environments. We conduct a comparative analysis of various trash classification methods utilizing deep learning models built on convolutional neural networks (CNNs). Leveraging the PyTorch open-source framework and the TrashBox dataset, we perform experiments involving ten unique deep neural network models. Our approach aims to maximize training accuracy. Through extensive experimentation, we observe the consistent superiority of the ResNext-101 model compared to others, achieving exceptional training, validation, and test accuracies. These findings illuminate the potential of CNN-based techniques in significantly advancing trash classification for optimized solid waste management within smart city initiatives. Lastly, this study presents a distributed framework based on federated learning that can be used to optimize the performance of a combination of CNN models for trash detection.

Analisis Bencana Banjir Menggunakan Data SAR (Synthetic Aperture Radar) Sentinel 1 di Kecamatan Sungai Serut Kota Bengkulu

Sungai Serut Sub-district is one of the sub-districts in Bengkulu City where flooding occurs most frequently. The objectives of this study were to 1) analyze flood disasters and determine the extent of the flood area on January 06, 2022 and February 06, 2022 in Sungai Serut Subdistrict, 2) determine the number of buildings affected by flooding in 2022 in Sungai Serut Subdistrict. The method used in this research is Change Detection and Thresholding and satellite image digitization, to assess the accuracy of the data in accordance with the situation in the field, an accuracy test analysis is carried out using overall accuracy. The results showed that 1) based on the results of the analysis, the flood area on January 06, 2022 was 15.5 hectares and on February 06, 2022 was 146.32 hectares, 2) the number of buildings affected by flooding in 2022 in Sungai Serut Subdistrict was 425 buildings spread across each urban village with different numbers of buildings. Surabaya Village is the area with the highest number of buildings affected by flooding, around 343 buildings. The level of data accuracy test results obtained an accuracy percentage of 97.11% for the overall accuracy value.

Clinicopathological and molecular characterization of resected lung adenocarcinoma: Correlations with histopathological grading systems in Chinese patients

PurposeDriver mutations inform lung adenocarcinoma (LUAD) targeted therapy. Association of histopathological attributes and molecular profiles facilitates clinically viable testing platforms. We assessed correlations between LUAD clinicopathological features, mutational landscapes, and two grading systems among Chinese cases. Methods79 Chinese LUAD patients undergoing resection were subjected to targeted sequencing. 68 were invasive nonmucinous adenocarcinoma (INMA), graded via: predominant histologic pattern-based grading system (P-GS) or novel IASLC grading system (I-GS). Driver mutation distributions were appraised and correlated with clinical and pathological data. ResultsCompared to INMA, non-INMA exhibited smaller, well-differentiated tumors with higher mucin content. INMA grade correlated with size, lymph invasion (P-GS), and driver/EGFR mutations. Mutational spectra varied markedly between grades, with EGFR p.L858R and exon 19 deletion mutations predominating in lower grades; while high-grade P-GS tumors often harbored EGFR copy number variants and complex alterations alongside wild-type cases. I-GS upgrade of P-GS grade 2 to grade 3 was underpinned by ≥20 % high-grade regions bearing p.L858R or ALK fusions. Both systems defined tumors of distinctive phenotypic attributes and molecular genotypes. ConclusionsINMA represent larger, mucin-poor, molecularly heterogeneous LUAD with divergent grade-specific mutation profiles. Stronger predictor of clinicopathological attributes and driver mutations, P-GS stratification offers greater accuracy for molecular testing. A small panel encompassing EGFR and ALK captures the majority of P-GS grade 1/2 mutations whereas expanded panels are optimal for grade 3.

Use of the Head Impulse, Nystagmus, Test of Skew ('HINTS') assessment to aid differential diagnosis in acute vestibular syndrome in the hyperacute stroke setting.

Differential diagnosis of acute vertigo syndrome is challenging given the similarities between clinical presentations of posterior circulation stroke and peripheral vestibular dysfunction. The Head Impulse, Nystagmus, Test of Skew ('HINTS') assessment is a clinical bedside test used to aid diagnosis. Comprehensive training on use of the Head Impulse, Nystagmus, Test of Skew assessment was provided to one stroke consultant, and the effectiveness of the test in that setting was assessed. Further education was completed with more members of the stroke and emergency department multi-disciplinary team. Quality improvement measures including magnetic resonance imaging use and bed utilisation were explored. Following training of one stroke consultant, the Head Impulse, Nystagmus, Test of Skew assessment was found to be a feasible, accurate bedside test within this acute stroke service. Further training for the multi-disciplinary team was completed, but outcome measures were not explored because of the coronavirus disease 2019 pandemic and maternity leave. There is a role for trained members of the multi-disciplinary team to successfully use the Head Impulse, Nystagmus, Test of Skew assessment in hyperacute stroke settings, to aid diagnosis in acute vertigo syndrome.

Integrating optical and electrical sensing with machine learning for advanced particle characterization

Particle classification plays a crucial role in various scientific and technological applications, such as differentiating between bacteria and viruses in healthcare applications or identifying and classifying cancer cells. This technique requires accurate and efficient analysis of particle properties. In this study, we investigated the integration of electrical and optical features through a multimodal approach for particle classification. Machine learning classifier algorithms were applied to evaluate the impact of combining these measurements. Our results demonstrate the superiority of the multimodal approach over analyzing electrical or optical features independently. We achieved an average test accuracy of 94.9% by integrating both modalities, compared to 66.4% for electrical features alone and 90.7% for optical features alone. This highlights the complementary nature of electrical and optical information and its potential for enhancing classification performance. By leveraging electrical sensing and optical imaging techniques, our multimodal approach provides deeper insights into particle properties and offers a more comprehensive understanding of complex biological systems.Graphical abstract

Helical Gearbox Defect Detection with Machine Learning Using Regular Mesh Components and Sidebands.

The current paper presents helical gearbox defect detection models built from raw vibration signals measured using a triaxial accelerometer. Gear faults, such as localized pitting, localized wear on helical pinion tooth flanks, and low lubricant level, are under observation for three rotating velocities of the actuator and three load levels at the speed reducer output. The emphasis is on the strong connection between the gear faults and the fundamental meshing frequency GMF, its harmonics, and the sidebands found in the vibration spectrum as an effect of the amplitude modulation (AM) and phase modulation (PM). Several sets of features representing powers on selected frequency bands or/and associated peak amplitudes from the vibration spectrum, and also, for comparison, time-domain and frequency-domain statistical feature sets, are proposed as predictors in the defect detection task. The best performing detection model, with a testing accuracy of 99.73%, is based on SVM (Support Vector Machine) with a cubic kernel, and the features used are the band powers associated with six GMF harmonics and two sideband pairs for all three accelerometer axes, regardless of the rotation velocities and the load levels.

Guidelines for Manufacturing and Application of Organoids: Liver.

Recent amendments to regulatory frameworks have placed a greater emphasis on the utilization of in vitro testing platforms for preclinical drug evaluations and toxicity assessments. This requires advanced tissue models capable of accurately replicating liver functions for drug efficacy and toxicity predictions. Liver organoids, derived from human cell sources, offer promise as a reliable platform for drug evaluation. However, there is a lack of standardized quality evaluation methods, which hinders their regulatory acceptance. This paper proposes comprehensive quality standards tailored for liver organoids, addressing cell source validation, organoid generation, and functional assessment. These guidelines aim to enhance reproducibility and accuracy in toxicity testing, thereby accelerating the adoption of organoids as a reliable alternative or complementary tool to animal testing in drug development. The quality standards include criteria for size, cellular composition, gene expression, and functional assays, thus ensuring a robust hepatotoxicity testing platform.

Assessment of gene–disease associations and recommendations for genetic testing for somatic variants in vascular anomalies by VASCERN-VASCA

BackgroundVascular anomalies caused by somatic (postzygotic) variants are clinically and genetically heterogeneous diseases with overlapping or distinct entities. The genetic knowledge in this field is rapidly growing, and genetic testing is now part of the diagnostic workup alongside the clinical, radiological and histopathological data. Nonetheless, access to genetic testing is still limited, and there is significant heterogeneity across the approaches used by the diagnostic laboratories, with direct consequences on test sensitivity and accuracy. The clinical utility of genetic testing is expected to increase progressively with improved theragnostics, which will be based on information about the efficacy and safety of the emerging drugs and future molecules. The aim of this study was to make recommendations for optimising and guiding the diagnostic genetic testing for somatic variants in patients with vascular malformations.ResultsPhysicians and lab specialists from 11 multidisciplinary European centres for vascular anomalies reviewed the genes identified to date as being involved in non-hereditary vascular malformations, evaluated gene–disease associations, and made recommendations about the technical aspects for identification of low-level mosaicism and variant interpretation. A core list of 24 genes were selected based on the current practices in the participating laboratories, the ISSVA classification and the literature. In total 45 gene–phenotype associations were evaluated: 16 were considered definitive, 16 strong, 3 moderate, 7 limited and 3 with no evidence.ConclusionsThis work provides a detailed evidence-based view of the gene–disease associations in the field of vascular malformations caused by somatic variants. Knowing both the gene–phenotype relationships and the strength of the associations greatly help laboratories in data interpretation and eventually in the clinical diagnosis. This study reflects the state of knowledge as of mid-2023 and will be regularly updated on the VASCERN-VASCA website (VASCERN-VASCA, https://vascern.eu/groupe/vascular-anomalies/).

Machine learning assisted classification between diabetic polyneuropathy and healthy subjects using plantar pressure and temperature data: a feasibility study

Automated and early detection of diabetics with polyneuropathy in an ambulatory health monitoring setup may reduce the major risk factors for diabetic patients. Increased and localized plantar pressure associated with impaired pain and temperature is a combination of developing foot ulcers in subjects with polyneuropathy. Although many interesting research works have been reported in this area, most of them emphasize on signal acquisition process and plantar pressure distribution in the foot region. In this work, a machine learning assisted low complexity technique was developed using plantar pressure and temperature signals which will classify between diabetic polyneuropathy and healthy subjects. Principal component analysis (PCA) and maximum relevance minimum redundancy (mRMR) methods were used for feature extraction and selection respectively followed by k-NN classifier for binary classification. The proposed technique was evaluated with 100 min of publicly available annotated data from 43 subjects and provides blind test accuracy, sensitivity, precision, F1-score, and area under curve (AUC) of 99.58%, 99.50%, 99.44%, 99.47% and 99.56% respectively. A low resource hardware implementation in ARM v6 controller required an average memory usage of 81.2 kB and latency of 1.31 s to process 9 s pressure and temperature data collected from 16 sensor channels for each of the foot region.

A comparative study of machine learning models for identifying noxious gases through thermal fingerprint measurements and MOS sensors

Recently, there has been a notable surge of interest in developing swift and precise gas detection and categorization tools through artificial intelligence. This work specifically focuses on designing a simple sensor array configuration and empirically compares various machine-learning models for identifying multiple gases. The study employs eight metal oxide semiconductor sensors, each operating independently based on the thermal fingerprint principle. These sensors are utilized to gather data on responses to different toxic gases, and a range of traditional machine learning classifiers are trained and assessed using this dataset. The classifiers undergo investigation across various train/test dataset split ratios and different principal component analysis dimensionalities. Average accuracy metrics are employed over the entire testing dataset to evaluate model performance. The results demonstrated the challenge of selectively detecting each gas type solely through conventional analyses with the gas response data of the proposed sensor array. Nevertheless, the utilization of machine learning models has exhibited promising outcomes in augmenting gas identification and classification. The findings reveal a consistent improvement in testing accuracy for the Gradient Boosting model, which achieves an outstanding accuracy rate, with instances of reaching 100% accuracy in the training dataset when three or more principal components are considered. This study highlights the potential of the proposed straightforward sensor array and presents effective recognition algorithms applicable to creating electronic noses for gas-sensing applications.

Classification of mango disease using ensemble convolutional neural network

Mango is a highly significant fruit crop that thrives in a variety of agro-ecologies around the world. Mangoes are rich in vitamins and minerals. However, its yield is currently severely constrained due to disease and pest infestations. Thus, in order to improve mango fruit quality and productivity, illnesses and insect pests must be detected early on. In this study, we conceived and constructed a mango leaf disease detection mechanism utilizing an ensemble convolutional neural network approach. Healthy and diseased mango leaf images were manually obtained from main producing locations in Amhara Region for Merawi fruit and vegetable research identification. To improve the datasets, several pre-processing procedures (such as image resizing, noise reduction, and image augmentation) were used. To improve classification performance and meet the study's purpose, various segmentation approaches such as k means and Mask R-CNN were applied. Furthermore, following pre-processing and segmentation, features of mango leaf images were retrieved using CNN to obtain important features. The classification model was then constructed using fully-connected layer classifiers on the retrieved features of mango leaf images. The ensemble proposed GoogLeNet and VGG16 based CNN model in the study encompasses various operations, including dataset collection, image preprocessing, noise removal, segmentation, data augmentation, feature extraction, and classification. Upon testing, the model demonstrated impressive performance with 99.87 % training classification accuracy, 99.72 % validation accuracy, and 99.21 % testing accuracy. This indicates the effectiveness of the ensemble approach in achieving high accuracy in image classification tasks.