Layer Count Research Articles

NAMES OF TREES AND SHRUBS IN A FEW POLISH DIALECTS IN UKRAINE

Flora is an inseparable element of rural life and, consequently, it preserves the linguistic image of the world in human consciousness. The paper presents the vocabulary relating to the world of trees and shrubs in a few Polish dialects in Khmelnytskyi and Zhytomyr regions in Ukraine based on several published dictionaries and unpublished texts. The examined thematic scope counts 149 lexical units, divided into five groups: “General notions and places of growing”, “Names of trees and shrubs”, “Parts of trees and shrubs”, “Characteristics of trees and shrubs” and “Functioning of trees and shrubs”. The analysis goes beyond presenting exclusively “peculiar” lexemes in favour of a comprehensive approach to the researched scope, what corresponds to the latest tendencies in studying Polish vocabulary in Ukraine. Thus, the article shows both native Polish lexical layer counting 84 words (i.e. 56%), of which 57 words (i.e. 68%) are common to Polish and Ukrainian and/or Russian, and borrowed layer counting 65 words from Ukrainian and/or Russian (i.e. 44%), among which Ukrainianisms definitely prevail (i.e. 68%). Showing the ratio of native Polish to borrowed lexical layer reveals, on the one hand, the strength of the native language, and on the other hand, the strength of foreign influences. Their competition manifests itself among others in broad synonymy that can appear on different levels. It is worth emphasizing that the research concerns word registers and dictionaries and consequently does not reflect the frequency of word use in texts. We suppose that a frequency study of thematic vocabulary would bring different results and should be the subject of further research.

A novel deep neural network model based Xception and genetic algorithm for detection of COVID-19 from X-ray images.

The coronavirus first appeared in China in 2019, and the World Health Organization (WHO) named it COVID-19. Then WHO announced this illness as a worldwide pandemic in March 2020. The number of cases, infections, and fatalities varied considerably worldwide. Because the main characteristic of COVID-19 is its rapid spread, doctors and specialists generally use PCR tests to detect the COVID-19 virus. As an alternative to PCR, X-ray images can help diagnose illness using artificial intelligence (AI). In medicine, AI is commonly employed. Convolutional neural networks (CNN) and deep learning models make it simple to extract information from images. Several options exist when creating a deep CNN. The possibilities include network depth, layer count, layer type, and parameters. In this paper, a novel Xception-based neural network is discovered using the genetic algorithm (GA). GA finds better alternative networks and parameters during iterations. The best network discovered with GA is tested on a COVID-19 X-ray image dataset. The results are compared with other networks and the results of papers in the literature. The novel network of this paper gives more successful results. The accuracy results are 0.996, 0.989, and 0.924 for two-class, three-class, and four-class datasets, respectively.

Exploring soluble and colloidally transported trace elements in stalagmites: The strontium-yttrium connection

While seasonality in speleothem trace element signatures is well-documented, the parameters that control the emergence of laminations vary between elements and tend to be multi-factorial. Here, we examine a series of active and fossil stalagmites from Asturias, Spain, with a particular focus on strontium and yttrium co-variations and fluorescent laminations. Coupled confocal fluorescence scanning light microscopy (layer counting) and time scales derived from accelerated mass spectrometry (F14C) in active stalagmites confirm that fluorescent banding is annual. This banding is coincident with Y peaks and Sr troughs, which are among the most robust trace element markers of seasonality. Strontium concentrations (in particular, the strontium partition coefficient, DSr) are positively correlated with stalagmite growth rate and are likely controlled by solution supersaturation, which is in turn controlled by seasonal variations in cave ventilation. DSr can be estimated after correcting for prior calcite precipitation using coeval Mg/Ca ratios, and is consistent with both empirical and experimental values. Meanwhile, yttrium is a proxy for colloidal organic input, and its concentration in stalagmites is likely controlled by a combination of Y drip water flux, surface retention time (i.e., how long a drip and its associated organic matter are in contact with the stalagmite surface), and dilation within the matrix (hereafter referred to as “dilation”). Persistent Sr-Y anti-correlation can be explained as an interplay between the individual controls on each element, and a breakdown in this relationship may be indicative of past changes in cave ventilation and/or drip hydrology.

Machine learning application to layer counting in speleothems

Some speleothems, in particular stalagmites, are laminated at the visible and microscopic scale, with the latter visible using fluorescence microscopy (e.g., confocal laser scanning microscopy). These laminations can be used to supplement speleothem chronologies, although this process is laborious and lateral variations in lamination geometry and quality necessitate a detailed look over an entire scan as opposed to a simple one-dimensional transect. In order to assist this process, we develop a classification-based machine learning algorithm using an open-source machine learning package. This algorithm is optimized for stalagmites growing at 20–100 μm yr−1 and outputs a 2-dimensional layer density map which may aid in quantitatively interpreting past variations in speleothem growth rate. This algorithm requires user supervision and interpretation, as image artefacts and magnification settings may complicate model output.

Predicting the Optimal Input Parameters for the Desired Print Quality Using Machine Learning.

3D printing is a growing technology being incorporated into almost every industry. Although it has obvious advantages, such as precision and less fabrication time, it has many shortcomings. Although several attempts were made to monitor the errors, many have not been able to thoroughly address them, like stringing, over-extrusion, layer shifting, and overheating. This paper proposes a study using machine learning to identify the optimal process parameters such as infill structure and density, material (ABS, PLA, Nylon, PVA, and PETG), wall and layer thickness, count, and temperature. The result thus obtained was used to train a machine learning algorithm. Four different network architectures (CNN, Resnet152, MobileNet, and Inception V3) were used to build the algorithm. The algorithm was able to predict the parameters for a given requirement. It was also able to detect any errors. The algorithm was trained to pause the print immediately in case of a mistake. Upon comparison, it was found that the algorithm built with Inception V3 achieved the best accuracy of 97%. The applications include saving the material from being wasted due to print time errors in the manufacturing industry.

Temporal and spatial laser intensification within nodular defects overcoated with multilayer dielectric mirrors over a wide range of defect geometries.

The nodular defect shape and the laser incidence angle have a dramatic impact on the spatial distribution of light intensification within the nodule as well as how the laser light is drained from the defect. Nodular defect geometries unique to ion beam sputtering, ion-assisted deposition, and electron-beam (e-beam) deposition, respectively, are modeled in this parametric study over a wide range of nodular inclusion diameters and layer count for optical interference mirror coatings constructed with quarter-wave thicknesses and capped with a half wave of the low index material. It was found for hafnia (n=1.9) and silica (n=1.45) multilayer mirrors that the light intensification in nodular defects with a C factor of 8, typical of e-beam deposited coatings deposited with a wide range of deposition angles, was maximized for a 24-layer design. For intermediate size inclusion diameters, increasing the layer count for normal incidence multilayer mirrors reduced the light intensification within the nodular defect. A second parametric study explored the impact of the nodule shape on the light intensification for a fixed number of layers. In this case, there is a strong temporal trend for the different nodule shapes. Narrow nodules tend to drain more laser energy through the bottom of the nodule into the substrate while wide nodules tend to drain more laser energy through the top of the nodule when irradiated at normal incidence. At a 45° incidence angle, waveguiding is an additional method to drain laser energy from the nodular defect. Finally, laser light resonates within nodular defects longer than within the adjacent nondefective multilayer structure.

Prefacing unexplored archives from Central Andean surface-to-bedrock ice cores through a multifaceted investigation of regional firn and ice core glaciochemistry

AbstractShallow firn cores, in addition to a near-basal ice core, were recovered in 2018 from the Quelccaya ice cap (5470 m a.s.l) in the Cordillera Vilcanota, Peru, and in 2017 from the Nevado Illimani glacier (6350 m a.s.l) in the Cordillera Real, Bolivia. The two sites are ~450 km apart. Despite meltwater percolation resulting from warming, particle-based trace element records (e.g. Fe, Mg, K) in the Quelccaya and Illimani shallow cores retain well-preserved signals. The firn core chronologies, established independently by annual layer counting, show a convincing overlap indicating the two records contain comparable signals and therefore capture similar regional scale climatology. Trace element records at a ~1–4 cm resolution provide past records of anthropogenic emissions, dust sources, volcanic emissions, evaporite salts and marine-sourced air masses. Using novel ultra-high-resolution (120 μm) laser technology, we identify annual layer thicknesses ranging from 0.3 to 0.8 cm in a section of 2000-year-old radiocarbon-dated near-basal ice which compared to the previous annual layer estimates suggests that Quelccaya ice cores drilled to bedrock may be older than previously suggested by depth-age models. With the information collected from this study in combination with past studies, we emphasize the importance of collecting new surface-to-bedrock ice cores from at least the Quelccaya ice cap, in particular, due to its projected disappearance as soon as the 2050s.

Use of δ18Oatm in dating a Tibetan ice core record of Holocene/Late Glacial climate

Ice cores from the northwestern Tibetan Plateau (NWTP) contain long records of regional climate variability, but refrozen meltwater and dust in these cores has hampered development of robust timescales. Here, we introduce an approach to dating the ice via the isotopic composition of atmospheric O2 in air bubbles (δ18Oatm), along with annual layer counting and radiocarbon dating. We provide a robust chronology for water isotope records (δ18Oice and d-excess) from three ice cores from the Guliya ice cap in the NWTP. The measurement of δ18Oatm, although common in polar ice core timescales, has rarely been used on ice cores from low-latitude, high-altitude glaciers due to (1) low air pressure, (2) the common presence of refrozen melt that adds dissolved gases and reduces the amount of air available for analysis, and (3) the respiratory consumption of molecular oxygen (O2) by micro-organisms in the ice, which fractionates the δ18O of O2 from the atmospheric value. Here, we make corrections for melt and respiration to address these complications. The resulting records of water isotopes from the Guliya ice cores reveal climatic variations over the last 15,000 y, the timings of which correspond to those observed in independently dated lake and speleothem records and confirm that the Guliya ice cap existed before the Holocene. The millennial-scale drivers of δ18Oice are complex and temporally variable; however, Guliya δ18Oice values since the mid-20th century are the highest since the beginning of the Holocene and have increased with regional air temperature.

In-situ monitoring of powder bed fusion of polymers using laser profilometry

Process monitoring allows for unique opportunities to capture information at intermediate layers during fabrication for part qualification, but also to improve process robustness by identifying issues before yield or quality is reduced. An industrial-grade laser profilometer has been integrated for the first time in a commercial polymer laser sintering machine and used to monitor powder bed quality and potentially qualify parts during production. Powder layer quality was measured in different conditions and changed drastically as a function of recoating speed, preheating temperature but also particle shape. The effective layer thickness was obtained for the first time in PBF of polymers, and gradually increased with the layer count until reaching a plateau at the nominal layer thickness. Powder layer density could be calculated as well and lies between 34 and 65 %, depending on the powder type and layer count. Finally, curling increased with decreasing preheating temperature, ranging from 70 µm to 350 µm i.e. until collision with the recoater. Based on the measured degree of curling, mitigation strategies could be integrated using a closed-loop feedback control. Effective process monitoring such as the one provided with laser profilometry may enable the development of new materials featuring more complex processing conditions, but also larger machines where thermal uniformity might be an issue.

Recoil and charged particle energy spectra from the natSi(n,x) reaction and the Si semiconductor detector response to the 14 MeV neutrons

The energy spectra of light and heavy charged particles generated by the 14 MeV neutrons in silicon nuclei and in the Si semiconductor detector (SSD) were analysed. The existing measurements were used to validate the corresponding energy differential cross sections from major evaluations ENDF/B-VIII.0 and JEFF-3.3. The required data were extracted by the various codes: NJOY21, SPKA and MCNP6. It was shown that both evaluations underestimate the high energy part of the 28Si(n,x)p- and α-particle spectra by ≤50%, whereas at smaller emission energies no experimental data even exist. The calculation of the atom displacement damage in SSD reveals that the secondary protons produce additionally ≈15% of damage, whereas the contribution from secondary alphas is impossible to estimate since the Si(α,x) recoil data are not available.The single experiment, carried out by K. Yageta and co-workers in 1988, reported the energy spectra in the Si semiconductor down to the rather low energy of ≈0.3 MeV, at which the production of reaction recoils dominates over the light ions. The Monte Carlo simulation of this experiment was performed by MCNP6. The generation and transport of all relevant ions were taken in account. The energy depositions from protons or alphas were summed with their associated recoils after correction for non-ionization losses. The neutron reactions in the dead layer of the SSD were shown to increase the counts in active layer by (1–20)% depending on sort of ions. The impact of the field funneling effect was studied. Consideration of all these underlying processes has eventually allowed a reasonable reproduction of this experiment.

Refining the metavariables of convolutional autoencoders for better performance on littoral acoustic backscattering data

Convolutional autoencoders have been used in an effort to learn latent encodings of littoral accoustic backscatter, but previous work did not explore a wide range of metavariable variation. Only the dimension of the encoding space, and the learning rate of the training optimizer were adjusted to achieve a functioning initial result. The other metavariables were assigned with some justification but optimality was not verified. We explore the effects of changing the method by which random weights are initialized as well as doubling the amount of random kernels per convolutional layer of the autoencoder. Additionally, we vary the layer count of the network, and the shapes of the kernels in the layers. Finally, various nonlinear activations are compared against the activation used initially. The results based on experimental field data from the TREX13 experiment will be presented. [This work is sponsored by ONR grant N00174-20-1-0016.]

Structure optimization of the protection inductor for the high energy density pulse pump power supply

High energy density pulse pump power supply is an important part of a large laser fusion facility. When a short-circuit fault occurs, massive energy injection will cause the capacitor and switch to damage or even explode. Therefore, it is of great significance to set a series protection inductor. In this paper, the structure of the widely used solenoid inductor is optimized. The geometric mean distance principle is introduced, and Neumann’s formula is extended to calculate the inductance of the solenoid. The electromagnetic force on each turn is solved by the virtual displacement method. The analytical expression of the electromagnetic stress in the current-carrying solenoid is deduced by combining the force balance equation, constitutive equation and geometric constraint equation. The electromagnetic-structure coupling simulation verifies the accuracy of the expression. Based on the Taguchi method, the structure parameters of the solenoid are optimized. The results show that the solenoid’s wire radius and layer count are the two most important influencing factors. As the wire radius increases and the layer count increases, the maximum carrying energy per unit volume increases. The influence of insulation encapsulation on the stress distribution in the inductor is studied based on electromagnetic-structure coupling simulation. It can be known that insulation encapsulation can effectively reduce stress and restrain deformation. On this basis, a novel type of double-layer solenoid inductor is designed and processed. Several pulse discharge tests verify its large current-carrying capacity and high reliability. The above results are useful for structure design and application of the protection inductor.

The ST22 chronology for the Skytrain Ice Rise ice core – Part 1: A stratigraphic chronology of the last 2000 years

Abstract. A new ice core was drilled in West Antarctica on Skytrain Ice Rise in field season 2018/2019. This 651 m ice core is one of the main targets of the WACSWAIN (WArm Climate Stability of the West Antarctic ice sheet in the last INterglacial) project. A present-day accumulation rate of 13.5 cm w.e. yr−1 was derived. Although the project mainly aims to investigate the last interglacial (115–130 ka), a robust chronology period covering the recent past is needed to constrain the age models for the deepest ice. Additionally, this time period is important for understanding current climatic changes in the West Antarctic region. Here, we present a stratigraphic chronology for the top 184.14 m of the Skytrain ice core based on absolute age tie points interpolated using annual layer counting encompassing the last 2000 years of climate history. Together with a model-based depth–age relationship of the deeper part of the ice core, this will form the ST22 chronology. The chemical composition, dust content, liquid conductivity, water isotope concentration and methane content of the whole core was analysed via continuous flow analysis (CFA) at the British Antarctic Survey. Annual layer counting was performed by manual counting of seasonal variations in mainly the sodium and calcium records. This counted chronology was informed and anchored by absolute age tie points, namely, the tritium peak (1965 CE) and six volcanic eruptions. Methane concentration variations were used to further constrain the counting error. A minimal error of ±1 year at the tie points was derived, accumulating to ± 5 %–10 % of the age in the unconstrained sections between tie points. This level of accuracy enables data interpretation on at least decadal timescales and provides a solid base for the dating of deeper ice, which is the second part of the chronology.

Application of visual stratigraphy from line-scan images to constrain chronology and melt features of a firn core from coastal Antarctica

AbstractEstablishing an accurate chronology is crucial for interpretation of ice core-based climatic records. While high snow accumulation rates characterise coastal Antarctica, thus enabling recovery of highly resolved climatic records, summertime melting at such low-elevation sites offers challenges in establishing a reliable chronological framework through traditional approaches using the seasonality of stable water isotope and ionic proxy records. Here, we assess visual stratigraphy (VS) obtained from line-scan images as a proxy for annual layer counting in firn section (top 50 m) of the IND-36/B9 ice core (dated 1919–2016 CE) from the Djupranen Ice Rise in central Dronning Maud Land, East Antarctica. We also used these images to obtain melt history for the site and found that traditional thickness-based quantification of melt proportion results in significant overestimations. Since density has dominant control on VS profile over the firn section, we first used circulant single-spectrum analysis to remove the secular trend and then we extracted the seasonal VS signals attributed to dust and sea-salt inclusions. We find that melt layers do not significantly alter the VS records if masked during pre-processing. The age–depth model based on the reconstructed VS profile revealed an excellent match with identified time-markers within an uncertainty of ±2 years.

CVD growth and optical characterization of homo and heterobilayer TMDs

The stacking of few layers of transition metal dichalcogenides (TMDs) and their heterostructures allows us to create new structures, observe new physical phenomena, and envision new applications. Moreover, the twist angle in few-layer TMDs can significantly impact their electrical and optical properties. Therefore, controlling the TMD material and obtaining different stacking orientations when synthesizing TMDs via chemical vapor deposition (CVD) is a powerful tool, which can add functionality to TMD-based optoelectronic devices. Here, we report on the synthesis of few-layer MoS2 and WS2 crystals, as well as their heterobilayer structures with 0° and 60° twist angles between layers via CVD. Raman and photoluminescence spectroscopies demonstrate the quality, crystallinity, and layer count of our grown samples, while second harmonic generation shows that adjacent layers grow with 0° or 60° twist angles, corresponding to two different crystal phases. Our study based on TMDs with different and multiple stacking configurations provides an alternative route for the development of future optoelectronic and nonlinear optical devices.

Ciprofloxacin-Loaded Titanium Nanotubes Coated with Chitosan: A Promising Formulation with Sustained Release and Enhanced Antibacterial Properties.

Due to their high entrapment efficiency, anodized titanium nanotubes (TiO2-NTs) are considered effective reservoirs for loading/releasing strong antibiotics whose systemic administration is associated with diverse and severe side-effects. In this study, TiO2-NTs were synthesized by anodic oxidation of titanium foils, and the effects of electrolyte percentage and viscosity on their dimensions were evaluated. It was found that as the water content increased from 15 to 30%, the wall thickness, length, and inner diameter of the NTs increase from 5.9 to 15.8 nm, 1.56 to 3.21 µm, and 59 to 84 nm, respectively. Ciprofloxacin, a highly potent antibiotic, was loaded into TiO2-NTs with a high encapsulation efficiency of 93%, followed by coating with different chitosan layers to achieve a sustained release profile. The prepared formulations were characterized by various techniques, such as scanning electron microscopy, differential scanning calorimetry, and contact measurement. In vitro release studies showed that the higher the chitosan layer count, the more sustained the release. Evaluation of antimicrobial activity of the formulation against two endodontic species from Peptostreptococcus and Fusobacterium revealed minimum inhibitory concentrations (MICs) of 1 µg/mL for the former and the latter. To summarize, this study demonstrated that TiO2-NTs are promising reservoirs for drug loading, and that the chitosan coating provides not only a sustained release profile, but also a synergistic antibacterial effect.

Comprehensive uncertainty estimation of the timing of Greenland warmings in the Greenland ice core records

Abstract. Paleoclimate proxy records have non-negligible uncertainties that arise from both the proxy measurement and the dating processes. Knowledge of the dating uncertainties is important for a rigorous propagation to further analyses, for example, for identification and dating of stadial–interstadial transitions in Greenland ice core records during glacial intervals, for comparing the variability in different proxy archives, and for model-data comparisons in general. In this study we develop a statistical framework to quantify and propagate dating uncertainties in layer counted proxy archives using the example of the Greenland Ice Core Chronology 2005 (GICC05). We express the number of layers per depth interval as the sum of a structured component that represents both underlying physical processes and biases in layer counting, described by a regression model, and a noise component that represents the fluctuations of the underlying physical processes, as well as unbiased counting errors. The joint dating uncertainties for all depths can then be described by a multivariate Gaussian process from which the chronology (such as the GICC05) can be sampled. We show how the effect of a potential counting bias can be incorporated in our framework. Furthermore we present refined estimates of the occurrence times of Dansgaard–Oeschger events evidenced in Greenland ice cores together with a complete uncertainty quantification of these timings.

A multi-ice-core, annual-layer-counted Greenland ice-core chronology for the last 3800 years: GICC21

Abstract. Ice-core timescales are vital for the understanding of past climate; hence they should be updated whenever significant amounts of new data become available. Here, the Greenland ice-core chronology GICC05 was revised for the last 3835 years by synchronizing six deep ice cores and three shallow ice cores from the central Greenland ice sheet. A new method was applied by combining automated counting of annual layers on multiple parallel proxies and manual fine-tuning. A layer counting bias was found in all ice cores because of site-specific signal disturbances; therefore the manual comparison of all ice cores was deemed necessary to increase timescale accuracy. After examining sources of error and their correlation lengths, the uncertainty rate was quantified to be 1 year per century. The new timescale is younger than GICC05 by about 13 years at 3835 years ago. The most recent 800 years are largely unaffected by the revision. Between 800 and 2000 years ago, the offset between timescales increases steadily, with the steepest offset occurring between 800 and 1100 years ago. Moreover, offset oscillations of about 5 years around the average are observed between 2500 and 3800 years ago. The non-linear offset behavior is attributed to previous mismatches of volcanic eruptions, to the much more extensive dataset available to this study, and to the finer resolution of the new ice-core ammonium matching. By analysis of the common variations in cosmogenic radionuclides, the new ice-core timescale is found to be in alignment with the IntCal20 curve (Reimer et al., 2020).

Quantity and Distribution of Eosinophils in the Adult Human Gastric Specimens.

Background & Objective:The existence of eosinophils in the gastric mucosal epithelium is unusual, while the term "eosinophilic gastritis" has become overused due to the increased numbers of eosinophils found in gastric specimens. Thus, we aimed to assess the number and distribution of eosinophils in gastric specimens in Shahid Beheshti University of Medical Sciences hospitals. Methods:This study was performed on 318 patients with gastric diseases who had undergone endoscopic biopsy or gastrectomy in the hospitals affiliated with Shahid Beheshti University from 2016 to 2018. By referring to the archives of pathology departments, patients' demographic and clinical information, endoscopic and histopathological findings were collected. The data was then statistically analyzed using SPSS software version 24 with a significance level of P-value< 0.05 in all tests.Results:The participants were 157 men and 161 women, with an average age of 51.21 years. There was no significant correlation between eosinophil distribution and age, gender, or primary gastric locations. However, there was a strong correlation between the count of eosinophils in the lamina propria and intestinal metaplasia. Mean eosinophil count per high power field (HPF) was 12, 23, and 14 in mild, moderate, and severe degrees of intestinal metaplasia, respectively. An increase in eosinophil count was seen along with lymphoplasma cells infiltration up to 8/HPF in mild cases, 13/HPF in moderate cases, and 14/HPF in severe cases. Conclusion: Eosinophil counts in the lamina propria layer show a significant positive relationship with the eosinophil sheet, presence of Heliobacter pylori microorganism, intestinal metaplasia, and lymphoplasma cells infiltration.

A Novel MRI Diagnosis Method for Brain Tumor Classification Based on CNN and Bayesian Optimization.

Brain tumor is one of the most aggressive diseases nowadays, resulting in a very short life span if it is diagnosed at an advanced stage. The treatment planning phase is thus essential for enhancing the quality of life for patients. The use of Magnetic Resonance Imaging (MRI) in the diagnosis of brain tumors is extremely widespread, but the manual interpretation of large amounts of images requires considerable effort and is prone to human errors. Hence, an automated method is necessary to identify the most common brain tumors. Convolutional Neural Network (CNN) architectures are successful in image classification due to their high layer count, which enables them to conceive the features effectively on their own. The tuning of CNN hyperparameters is critical in every dataset since it has a significant impact on the efficiency of the training model. Given the high dimensionality and complexity of the data, manual hyperparameter tuning would take an inordinate amount of time, with the possibility of failing to identify the optimal hyperparameters. In this paper, we proposed a Bayesian Optimization-based efficient hyperparameter optimization technique for CNN. This method was evaluated by classifying 3064 T-1-weighted CE-MRI images into three types of brain tumors (Glioma, Meningioma, and Pituitary). Based on Transfer Learning, the performance of five well-recognized deep pre-trained models is compared with that of the optimized CNN. After using Bayesian Optimization, our CNN was able to attain 98.70% validation accuracy at best without data augmentation or cropping lesion techniques, while VGG16, VGG19, ResNet50, InceptionV3, and DenseNet201 achieved 97.08%, 96.43%, 89.29%, 92.86%, and 94.81% validation accuracy, respectively. Moreover, the proposed model outperforms state-of-the-art methods on the CE-MRI dataset, demonstrating the feasibility of automating hyperparameter optimization.