Small Section Research Articles

Centrifuge model studies on the performance of earthen dams with geocomposite internal drain subjected to pseudostatic seismic loading

This paper evaluates geocomposite (GC) as an internal drain for seepage control in earthen dams subjected to upstream impoundment and inertial loading through centrifuge model tests. A total of five centrifuge tests were performed on a homogeneously built small earthen dam section having prototype height of 7.2 m with 1H:1V side slopes at 30 g-level using a 4.5 m radius large beam geotechnical centrifuge available at the Indian Institute of Technology Bombay, India. The studied earthen dam models had an internal drain in the form of (a) a geocomposite horizontal drain (GC HD); (b) a sand-sandwiched geocomposite horizontal drain (SSGC HD); (c) a geocomposite chimney drain (GC CD); and (d) a sand-sandwiched geocomposite chimney drain (SSGC CD). The seepage-deformation response of the GC equipped earthen dam models is presented in terms of the development of pore-water pressure (PWP), phreatic surfaces, surface settlements (sc), and downstream slope face movements (sf). The behavior of the earthen dam model without any internal drain is also presented for comparison. In the case of an earthen dam with GC internal drain both as HD and CD, it was found to be efficient up to some extent in restricting phreatic surfaces, but phreatic surfaces were observed to be close to downstream side slope. However, a sand-sandwiched GC internal drain HD and CD seemed more promising in restricting the phreatic surface much within the downstream region and imparting higher stability even during sustained inertial loading conditions. The study concludes that SSGC chimney drain can be safely adopted for seepage control in small earthen dams and similar geotechnical structures viz levees, raw water reservoirs (RWRs), coastal protection structures, etc.

River discharge recovery lag in the small and medium-sized rivers occurred in response to climate change and human activities

To cope with the negative impacts of climate change and human activities on water resources in basins, positive measures such as interbasin water transfer projects and efficient water resource policies have been implemented worldwide, and these measures have had an obvious effect on the discharges of large rivers. However, for small and medium-sized rivers, the impact is still unclear due to a deficiency in river discharge data. To answer this question, the Haihe River Basin (HRB), located in the core area of economic development in North China, was selected as a typical research area. Remote sensing hydrological station technology, which combines unmanned aerial vehicles (UAVs), satellite remote sensing, and in situ surveying, was applied to obtain the monthly scale discharges of 16 small and medium-sized river sections in the last 20 years (2000–2020). In addition, an accuracy evaluation and spatiotemporal change analysis of the discharge results, which were influenced by interbasin water transfer projects and efficient water resource management policies, were also carried out. The results show that 1) 10 of the 16 river sections in the basin had a downward trend in the past 20 years, with an average annual decrease of 2.13%, and these sections were located in the plain areas; the remaining 6 sections located in the mountain areas had an upward trend, with an average annual increase of 1.72%; 2) we found that the river discharges of small and medium-sized rivers recovered with a lag that reached 3–11 years for the 16 rivers, and this lag for the mountain rivers was 3–9 years and for the plain rivers, it was 8–11 years; and 3) the main cause of this lag was the heavy exploitation of groundwater in the basin, which led to increased water volume from interbasin water transfers mainly to supplement groundwater resources, while surface water resources could not be quickly and sufficiently supplemented to restore them to the original natural discharge. Our research not only unmasks the changes of small and medium-sized river discharge in areas strongly affected by human activities but also helps people to better understand the response of small and medium-sized river discharges to the complex environmental changes incurred by climate change and serious human activities.

A point cloud preprocessing method for complicated thin-walled ring parts with irregular section

Metallic W-ring is a typically complicated thin-walled ring part with irregular section (CTWRPIC). However, there is a lack of effective measurement methods to inspect its geometric quality in the forming process, which leads to the failure of finding the forming defects in time. Due to the smooth metal surface, complex section shape, small section size and annular closed overall structure, point cloud of the section profile obtained by the line laser scanners has serious noise, outliers and missing. To this problem, this article proposes a point cloud preprocessing method for CTWRPIC. In this method, historical measurement data based outliers statistics, initial-fine rigid registration and non-rigid registration of CAD model are introduced. Firstly, the optimal alignment of point cloud from the CAD model and measurement point cloud is realized by initial-fine rigid registration. Secondly, the point cloud of CAD model is transformed by non-rigid registration to approximate the measurement profile points, the transformed point cloud of CAD model is considered as the real profile. Thirdly, based on the transformed point cloud of CAD model and distance threshold, the adaptive removal of outliers and repair of missing are realized. Finally, the correction of profile point cloud noise is completed by combining the Alpha Shapes algorithm and the proposed repair method. Based on the proposed preprocessing method and a measurement platform equipped with line laser scanners, an on-machine measurement method for metallic W-ring is proposed, and experimental results demonstrate the advantage of the proposed method. Compared with the destructive inspection method, the mean value, standard deviation and root mean square of the distance deviation of the preprocessed point cloud are all within 11 μm.

Resident-teaching of ethics in undergraduate medical education: a grounded theory analysis

Background In graduate medical education, teaching is a required subcompetency largely fulfilled via clinical teaching, journal clubs, and grand rounds. Evidence shows that when moving to undergraduate teaching, residents often face a steep learning curve. We aimed to assess residents’ perspective of the experience of teaching medical students. Methods Psychiatry residents taught small group sections of bioethics to first- and second-year medical students in December 2018. We conducted two 1-hour focus group interviews with four residents on their perspectives on the teaching experience. Results Resident-teachers described receiving certain benefits from teaching, such as meeting their altruistic desire to give back to the profession. Nonetheless, some participants felt frustrated by students’ varying engagement and respect, while also feeling insecure and intimidated. Resident-teachers experienced some of the medical students as disrespectful and limited in their appreciation for diversity and the profession of medicine and perceived the students’ disengagement and lessened professionalism. Conclusion As residency programs seek to implement initiatives to improve teaching skills of residents, resident experiences should be considered when implementing these initiatives.

Functionality analysis of special active ingredients from skincare products: Seaweed extract, Yeast ferment filtrates and Probiotics

This paper is a comparative study for three of the most popular ingredients used within different skincare products for particular conditioning purposes. Those of which include seaweed and kelp extracts (phlorotannins) and yeast ferment filtrates. A special case for a newly discovered treatment--topical probiotic use--will also be discussed. The origin and use, as well as the experimental effectiveness of these ingredients will be examined and introduced in the paper. Main skin conditions each one of them targets will be briefly discussed in small sections. The paper will also examine the benefits of each ingredient as well as their advantage over other products. Concerns around the topics of new and developing skincare formulas, as well as their ingredient’s extraction, will be discussed. Such concerns include, but are not limited to, environmental damage, ethical conflicts, social controversy.

Analysis of the Effect of Pore Water Pressure on a Small Radius Curve Section of a Fine Sand Layer under Cyclic Metro

Small curved metro shield tunnels located in fine sand layers are sensitive to the response of horizontal and vertical cyclic loads from train operations, especially for centrifugal horizontal loads. The majority of Zhengzhou’s strata are dominated by this geological composition. Therefore, the dynamic response of the fine sand layer under the train vibration load will lead to the settlement of the sand layer, which brings great hidden danger to the train operation. Long-term pore water monitoring was carried out in this paper, and the use of MIDAS-GTS (Multi-candidate Iterative Design with Adaptive Selection) finite element calculation platform to establish the metro ballast-lining-soil coupling dynamic model for mutual verification. The variation patterns of pore water pressure and super pore water pressure during train operation and the vibration response pattern of the soil layer around the tunnel were investigated. The results suggest that: (1) The pore and excess pore water pressures generated at the start of vibration are not easily dissipated and transferred, making them larger in the early stages of train operation. In contrast, the fine-grained powdered sandy soil has a small amount of clay particles, giving strength and cohesion to the soil layer. Vibrating hole pressure and excess pore water pressure stabilize with the train at a later stage; (2) The low probability of liquefaction in the silt layer surrounding the tunnel; (3) Under vibrating loads, areas of significant soil settlement are concentrated on the soil surface, on the upper side of the tunnel in the silty sand layer and at the bottom 3 m of the tunnel, however, its low variation in settlement has a low impact on the tunnel.

Plasmonic Metasurface Integrated Black Phosphorus‐Based Mid‐Infrared Photodetector with High Responsivity and Speed

AbstractBlack Phosphorus (BP), a van der Waals (vdW) semiconducting material, is studied intensively due to its unique structural and anisotropic optical properties. BP exhibits optical anisotropy with thickness tunable direct bandgap from 0.3–2.0 eV. Many BP‐based photodetectors operating in the mid‐infrared (MIR) band are proposed, but their performance is still limited due to the small optical absorption cross section at room temperature. In this work, a plasmonic metasurface that has localized surface plasmon resonance (LSPR) at 3.7 µm is designed, matching well with the band edge of a BP flake and field confinement around the edge of Au‐disk as confirmed by finite difference time domain simulation and Fourier‐transform infrared spectroscopy measurement results. In addition, a BP flake is integrated on the plasmonic metasurface and a significant quenching (twelve‐fold) is observed in the photoluminescence owing to the Förster resonance energy transfer effect induced by dipole–dipole coupling. Following that, a BP‐based MIR photodetector is fabricated on the plasmonic metasurface. Here, the demonstrated BP/plasmonic metasurface‐based photodetector achieves the peak responsivity of 495.85 mAW−1 and ultrahigh operation speed (>10 MHz) at the power of 8.55 µW under 3.7 µm incident wavelength. This demonstrated photodetector opens a new opportunity for optoelectronic applications in the MIR region.

A full-resolution convolutional network with a dynamic graph cut algorithm for skin cancer classification and detection

A robust medical decision support system for classifying skin lesions from dermoscopy images is a crucial instrument for determining skin cancer prognosis. In recent years, full resolution convolutional network has made significant progress in recognizing skin cancer types from Dermoscopic images despite their fine-grained changes in appearance. Recently, full resolution convolutional network have gained popularity as a solution to semantic segmentation issues. However, the hyper-parameters it chooses are what determine how well it performs, and manually fine-tuning these hyper-parameters takes time. Therefore, a hyper-parameter optimized full resolution convolutional network is suggested for dermoscopy picture segmentation in this research. The network’s hyper-parameters are optimized by a brand-new dynamic graph cut algorithm method. Hyper-parameters emphasize the proper balance between exploration and exploitation by combining the wolves’ individual haunting tactics with their global haunting strategies to generate a neighborhood-based searching strategy. The fundamental objective of this study is to develop a hyper-parameter-optimized Full resolution convolutional network-based model capable of reliably diagnosing skin cancer types using dermoscopy images. The computer-aided diagnosis could be more efficient and precise. The segmentation approach is the primary way to identify cancerous tumors with precision. This study introduces a dynamic graph cut algorithm -based method for accurate segmentation and improved skin cancer classification using a full resolution convolutional network. Experiments reveal that the proposed model effectively addresses the frequent over-segmentation and under-segmentation issues in graph cut and the subject of wrongly segmented small sections in the grab cut method. In addition, the results illustrate the utility of data augmentation in training and testing, with enhanced performance compared to the usage of fresh images. Multiple experiments were done using various transferring models, and the results of our recommended model showed superior performance in skin lesion categorization tasks relative to other architectures with an accuracy of 97.986%.

Accent rating of noise- and tone-vocoded foreign-accented speech

There are abundant acoustic-phonetic cues in speech signals for listeners to encode talker identity. However, speech signals in the real world are always less optimal due to various adverse listening sources. Vocoded speech is one type of simplified signal that has less spectral and/or temporal information in comparison to normal speech. The purpose of this study is to examine whether and how listeners’ judgment of talker accent is affected by noise and tone vocoding. Twelve Mandarin-accented English speakers with varying degree of accentedness and two native English speakers were recorded reading “The Rainbow Passage.” The recorded speech samples from each talker were segmented into small sections that were randomly selected for noise- or tone-excited vocoder processing into 1, 2, 4, 8, and 16 channels. The vocoded and unprocessed speech samples were randomly presented to a group of normal-hearing, monolingual English listeners for accent rating. The listeners judged the degree of talker accent on a 9-point Likert scale with “1” representing no accent and “9” representing extremely strong accent. The data are still in the process of being collected and analyzed. Results and implications of the present study will be discussed.

Enhanced Thomson scattering x-ray sources with flying focus laser pulse

X-ray source based on the Thomson scattering of laser interacting with energetic electron beams features high photon energy, small spot size, and good collimation. However, the photon number is insufficient for practical application because of the small cross section of the Thomson scattering. To solve this problem, here, we replace a traditional Gaussian laser pulse with a flying focus laser pulse to extend interaction length and restrain nonlinear effects. Simulation results show that the scattered photon number can be increased by about 25 and 2 times for high and low energy lasers, respectively. In particular, a 1010 photon number can be generated with a 10 J flying focus laser pulse, and the energy spread can also be greatly reduced for high energy laser, from a broad spectrum to a monoenergetic peak. Combining these two advantages, the peak spectrum brightness of x ray is 3 × 108 photons/keV at 240 keV, which is about three orders of magnitude more than the traditional case.

Poultry By-Product Meal as a Potential Protein Source in Broiler Diets with Exogenous Protease Supplementation

The objective of this study was to investigate the effect of exogenous protease supplementation in diets formulated with poultry by-product meal on growth performance, small intestine magnitudes, and meat physiochemical characters in broiler chickens from 21 to 35 days post-hatch. A total of 120, one-day-old "Arbor Acres Plus" broiler chickens (male: female - 50:50) were allocated one of two dietary treatments to give six replicates and ten birds per cage. Two dietary treatments included a control diet (CON) and a diet supplemented with protease (CON+Pro). At day 35, body weight and feed intakes were measured to calculate the feed conversion ratio for the entire experiment period. Two birds from each pen were euthanized to measure the relative lengths and relative weights of three small intestine sections and meat samples were collected for physiochemical characteristic analyses at 35 days post-hatch. Exogenous protease supplementation did not influence (<i>P</i>&gt;0.05) growth performance but showed a tendency to improve FCR (<i>P</i>=0.082). Protease supplementation showed a tendency to reduce proximal small intestine length (<i>P</i>=0.091). Broilers offered dietary treatments influenced minced meat color where protease supplementation resulted in lower CIE b^* (<i>P</i>&lt;0.001) colorimetric value for yellowness and showed a significant trend (<i>P</i>=0.059) on reducing meat redness CIE a^*. In conclusion, the addition of exogenous protease to a broiler diet formulated with poultry by-product meal did not affect the growth performance, small intestine magnitudes, and meat physiochemical parameters (except CIE b^*) in broiler chickens.

Millimeter-Wave Patch Antennas With Ultralow Profile of 20 μm on Flexible Substrate

In applications of flexible communication electronics, the requirement for ultralow profile of antenna is gradually approaching a limit. In order to achieve a higher flexibility, it is necessitated to limit the vertical distance to 10 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu \text{m}$ </tex-math></inline-formula> scale between the integrated antennas and the electrode ground (GND). Such a small Section height results in a severely reduced radiation efficiency and a narrowed bandwidth. In this article, the enhancement of radiation efficiency and bandwidth of a millimeter-wave (mmWave) patch antenna with 20 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu \text{m}$ </tex-math></inline-formula> profile height are studied. On the metal GND, by properly opening two (or one) resonant slots with a width of 100 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu \text{m}$ </tex-math></inline-formula> , strong resonance coupling between the patch antenna and the slots can be induced, leading to a greatly enhanced radiation efficiency. Meanwhile, the spectra width of TM10 mode of the patch antenna can also be effectively broadened by the mode coupling, giving rise to a greatly improved radiation bandwidth. The physical mechanism and the sensitive factors to maximize bandwidth, antenna gain, and front-to-back ratio are discussed.

Influence of different track structures on the development of corrugation in small radius curve section of subway

Many studies have shown that the rail corrugation on the small radius curve section of subway floating slab track is serious. And it seems to be different from the rail corrugation on the common track. The track structure must play an important role in the generation and development of rail corrugation. This paper measured a subway line in Tianjin, including 2 different track structures, and analyzed the measured data by using the international evaluation and acceptance indexes of rail corrugation. The results show that the inner rail of short sleeper monolithic track (SSMT) has poor smoothness in the wavelength range of 0-300 mm, and the overall smoothness of outer rail is good; the inner rail of steel spring floating slab track (SSFST) has poor smoothness in the wavelength range of 100-300 mm, and the outer rail has poor smoothness in the wavelength range of 300-1000 mm. Then the vehicle-track coupling model and wheel-rail wear calculation model is established. The calculation results show that in the existing rail corrugation, the short-wavelength rail corrugation of 20-25 mm will continue to appear or deteriorate, and the development of rail corrugation of other wavelengths will slow down or stop to varying degrees; with the increase of train passing times, the rail corrugation of SSFST with the wavelength of 315 mm develops the fastest, and the rail corrugation of SSMT with wavelengths of 20 mm and 160 mm develops the fastest.

Research on 3-D Precise Mapping System for Deformation Defects of Submarine Pipeline

Abstract The submarine oil pipeline has many advantages, such as large oil transportation capacity, and being fast and economical. However, long-term laid submarine oil pipelines are affected by reciprocating load of water flow, subsidence caused by soil liquefaction, ship anchorage operation, etc. In severe cases, it causes overall distortion of a small section of the submarine pipeline, profoundly affecting the safety of the submarine pipeline, which is a significant safety hazard for the health of the marine environment and potentially impacting social and economic benefits. Taking the Cezi-Zhenhai submarine pipeline in the sea area between Ningbo and Zhoushan as an example, many deformation defects in the pipeline have been found through internal inspection, and there is a trend of further deterioration. However, the existing external detection of submarine pipeline deformation can only collect limited data through mechanical dots for rough inversion. This does not meet the accuracy requirements of repairing existing submarine pipeline clamps. Therefore, we propose a real-time visualization surveying and mapping system for the submarine pipeline based on a 3-D laser and a separately designed electronic control system. Our research team performed actual mapping work for the Cezi-Zhenhai submarine pipeline and achieved an excellent mapping control effect. A steady monitoring image and good control effect of moving parts are obtained, and the data obtained by 3-D laser processing perfectly represent the actual state of the submarine pipeline. Predictably, the large-scale application of this system will provide a solid technical guarantee for the health of submarine pipelines.

Alternanthera philoxeroides extract as a corrosion inhibitor for steel in Cl3CCOOH solution

The corrosion inhibition of Alternanthera philoxeroides extract (APE) for cold rolled steel (CRS) in Cl3CCOOH solution was examined by weight loss and electrochemical measurements. AFM, SEM, CLSM and XPS techniques were executed to determine the inhibition mechanism. The results show that APE acts as an excellent inhibitor for CRS in 0.10 M Cl3CCOOH solution with the inhibition efficiency (ηw) of 94% at 200 mg L-1. The adsorption of APE on CRS surface follows Langmuir model at 20 and 30 oC, while Freundlich isotherm at 40 and 50 oC. APE is a mixed type inhibitor that mainly retards the anodic reaction. Nyquist diagram presents a depression capacitive loop at high frequency region followed by a small inductive section at low frequency region. With the addition of APE, the charge transfer resistance is significantly increased, but the double layer capacitance is dropped to some extent. Electrical conductivity further confirms that the presence of APE greatly hinders the corrosion of the metal by the acid medium. The efficient inhibition of APE is clearly confirmed from SEM and AFM micrographs. CLSM observation confirms the adsorption heterogeneous of APE. Meanwhile, XPS demonstrates that the polar groups of organic molecules in APE are bound to the iron matrix.

Supervised Contrastive Learning for Interpretable Long-Form Document Matching

Recent advancements in deep learning techniques have transformed the area of semantic text matching (STM). However, most state-of-the-art models are designed to operate with short documents such as tweets, user reviews, comments, and so on. These models have fundamental limitations when applied to long-form documents such as scientific papers, legal documents, and patents. When handling such long documents, there are three primary challenges: (i) the presence of different contexts for the same word throughout the document, (ii) small sections of contextually similar text between two documents, but dissimilar text in the remaining parts (this defies the basic understanding of “similarity”), and (iii) the coarse nature of a single global similarity measure which fails to capture the heterogeneity of the document content. In this article, we describe CoLDE : Co ntrastive L ong D ocument E ncoder—a transformer-based framework that addresses these challenges and allows for interpretable comparisons of long documents. CoLDE uses unique positional embeddings and a multi-headed chunkwise attention layer in conjunction with a supervised contrastive learning framework to capture similarity at three different levels: (i) high-level similarity scores between a pair of documents, (ii) similarity scores between different sections within and across documents, and (iii) similarity scores between different chunks in the same document and across other documents. These fine-grained similarity scores aid in better interpretability. We evaluate CoLDE on three long document datasets namely, ACL Anthology publications, Wikipedia articles, and USPTO patents. Besides outperforming the state-of-the-art methods on the document matching task, CoLDE is also robust to changes in document length and text perturbations and provides interpretable results. The code for the proposed model is publicly available at https://github.com/InterDigitalInc/CoLDE .

Feasibility Studies for the Measurement of Open-Charm Mesons at ALICE-3 Using Decay Channels with Neutral Mesons and Photons in the Final State

ALICE-3 is being designed as a next-generation heavy-ion experiment to be operated at the high-luminosity Large Hadron Collider. With luminosities higher by a factor of fifty, ALICE-3 will be able to study5 properties of quark–gluon matter with probes and precision which were previously unavailable due to small cross sections, high background levels, and insufficient detector sensitivity. In particular, the properties of hot and dense QCD matter will be studied by measuring production cross sections, flow coefficients, azimuthal angular correlations and nuclear modification factors for open-charm hadrons. In this contribution, we present the results of feasibility studies for the measurement of ground and excited states of open-charm mesons in decay channels D0 → K− + π+ + π0, D*(2007)0 → D0 + γ and D*(2010)± → D0 + π± in pp, p-Pb and Pb-Pb collisions at LHC energies using the ALICE-3 experimental setup. We formulate the main requirements for the selection of particles and their combinations to ensure reliable signal extraction in a wide transverse momentum range and estimate the minimum size of the required data samples. The results obtained are also compared to previous findings for the open-charm measurements in different decay channels.

Orbital evolution of satellite galaxies in self-interacting dark matter models

Dark matter self-interactions can leave distinctive signatures on the properties of satellite galaxies around Milky Way--like hosts through their impact on tidal stripping, ram pressure, and gravothermal collapse. We delineate the regions of self-interacting dark matter parameter space---specified by interaction cross section and a velocity scale---where each of these effects dominates and show how the relative mass loss depends on the satellite's initial mass, density profile, and orbit. We obtain novel, conservative constraints in this parameter space using Milky Way satellite galaxies with notably high central densities and small pericenter distances. Our results for self-interacting dark matter models, in combination with constraints from clusters of galaxies, favor either velocity-dependent cross sections that lead to gravothermal core collapse in the densest satellites or small cross sections which more closely resemble cold and collisionless dark matter.

Detection of acoustic emission from nanofiber nonwovens under tensile strain – An ultrasonic test setup for critical medical device components

In the biomedical field, nanofiber materials are gaining increasing application. For material characterization of nanofiber fabrics, tensile testing and scanning electron microscopy (SEM) are established standards. However, tensile tests provide information about the entire sample without information about single fibers. Conversely, SEM images examine individual fibers, but cover only a small section near the surface of the sample. To gain information on failure at the fiber level under tensile stress, recording of acoustic emission (AE) is a promising method, but challenging due to weak signal intensity. Using AE recording, beneficial findings can be obtained even on "invisible" material failure without affecting tensile tests. In this work, a technology for recording weak ultrasonic AE of tearing nanofiber nonwovens is presented, which uses a highly sensitive sensor. Functional proof of the method using biodegradable PLLA nonwoven fabrics is provided. The potential benefit is demonstrated by unmasking significant AE intensity in an almost imperceptible bend in the stress-strain curve of a nonwoven fabric. AE recording has not yet been performed on standard tensile tests of unembedded nanofiber material intended for safety-related medical applications. The technology has the potential to enrich the spectrum of testing methods, even those not confined to medical field.

Comparison of four convolutional neural networks for histopathological diagnosis of salivary gland carcinomas

Malignant salivary gland tumors represent aparticular diagnostic challenge due to the large number of histopathological entities, their rare occurrence, and the diverse clinical and histological presentations. The aim of this work is to investigate and compare convolutional neural networks (CNNs) as adiagnostic tool for histological diagnosis of salivary gland cancer. From salivary gland cancer preparations of 68patients, 118histological slides were digitized at high resolution. These virtual sections were then divided into small image sections, and the resultant 83,819images were sorted into four categories: background, connective tissue, non-neoplastic salivary gland tissue, and salivary gland cancer tissue. The latter category grouped the entities adenoid cystic carcinoma, adenocarcinoma (not otherwise specified), acinar cell carcinoma, basal cell carcinoma, mucoepidermoid carcinoma, and myoepithelial carcinoma. The categorized images were then processed in atraining, validation, and test run by the ImageNet pretrained CNN frameworks (Inception ResNetv2, Inceptionv3, ResNet152, Xception) in different pixel sizes. Accuracy values ranged from 18.8% to 84.7% across all network architectures and pixel sizes, with the Inceptionv3 network achieving the highest value at 500 × 500pixels. The recall values/sensitivity reached up to 85% for different pixel sizes (Inceptionv3 network at 1000 × 1000pixels). The minimum F1 score achieved was 0.07 for the InceptionResNetv2 and the Inceptionv3 at 100 × 100 pixels each, the maximum F1 score achieved was 0.72 for the Xception at 1000 × 1000pixels. Inceptionv3 was the network with the shortest training times, and was superior to all other networks at any pixel size. The current work was able to demonstrate the applicability of CNNs for histopathological analysis of salivary gland tumors for the first time and provide acomparison of the performance of different network architectures. The results indicate aclear potential benefit for future applications.