Article
Oncology
Yixiao Liu, Shen Jin, Qi Shen, Lufan Chang, Shancheng Fang, Yu Fan, Hao Peng, Wei Yu
Summary: This research demonstrates the effectiveness of a deep learning system in predicting the malignancy of urine cytology based on histopathology results. The system shows high sensitivity and specificity, providing novel insights for urologists in histopathology.
FRONTIERS IN ONCOLOGY
(2022)
Article
Biology
Vipin Venugopal, Justin Joseph, M. Vipin Das, Malaya Kumar Nath
Summary: This paper proposes a deep learning model, called DTP-Net, to address the thresholding problem of lesion localization on macro-images in dermatology. DTP-Net learns the intensity difference between the lesion and background and accurately predicts the threshold that separates them.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Computer Science, Information Systems
Shaohui Mei, Yunhao Geng, Junhui Hou, Qian Du
Summary: This paper proposes a method for reconstructing hyperspectral images (HSIs) using convolutional neural network (CNN) and easily acquired RGB images. The proposed SSR-Net can predict HSIs from RGB images without prior knowledge, and outperforms traditional methods in terms of HSI quality and classification performance.
SCIENCE CHINA-INFORMATION SCIENCES
(2022)
Article
Computer Science, Interdisciplinary Applications
Maria Chiara Fiorentino, Sara Moccia, Morris Capparuccini, Sara Giamberini, Emanuele Frontoni
Summary: This study proposed a framework based on deep learning methods for delineating fetal head circumference length through regression CNNs, achieving good results. Experimental results demonstrated the effectiveness of the framework in supporting clinicians during clinical practice.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2021)
Article
Genetics & Heredity
Julie E. Hoover-Fong, Kerry J. Schulze, Adekemi Y. Alade, Michael B. Bober, Ethan Gough, S. Shahrukh Hashmi, Jacqueline T. Hecht, Janet M. Legare, Mary Ellen Little, Peggy Modaff, Richard M. Pauli, David F. Rodriguez-Buritica, Maria E. Serna, Cory Smid, Chengxin Liu, John McGready
Summary: Achondroplasia is a common genetic skeletal disorder leading to short stature, and the CLARITY study aims to develop growth curves and reference tables for patients. With a large dataset collected, new growth curves have been constructed to enhance understanding of the clinical characteristics of patients with achondroplasia.
ORPHANET JOURNAL OF RARE DISEASES
(2021)
Article
Orthopedics
Xianghong Meng, Zhi Wang, Xinlong Ma, Xiaoming Liu, Hong Ji, Jie-zhi Cheng, Pei Dong
Summary: A deep convolutional neural network system was developed to automatically measure several parameters of the lower limbs. The system showed high consistency and accuracy in predicting the parameters, and its measurement time was significantly shorter than that of experienced radiologists.
BMC MUSCULOSKELETAL DISORDERS
(2022)
Article
Chemistry, Multidisciplinary
Suheng Xu, Alexander S. McLeod, Xinzhong Chen, Daniel J. Rizzo, Bjarke S. Jessen, Ziheng Yao, Zhicai Wang, Zhiyuan Sun, Sara Shabani, Abhay N. Pasupathy, Andrew J. Millis, Cory R. Dean, James C. Hone, Mengkun Liu, D. N. Basov
Summary: The study applies deep learning to analyze polaritonic waves in nanoscale images, utilizing a convolutional neural network to rapidly regress images and quantify the characteristics and material parameters of the waves. The CNN-based analysis method was validated through experiments and provides a general framework for quantitative information extraction.
Article
Chemistry, Multidisciplinary
Chang-bok Lee, Han-sung Lee, Hyun-chong Cho
Summary: Weight information is crucial in the breeding of cattle as it can assess animal growth and help calculate the appropriate daily feed amount. We have developed a non-invasive image-based method for estimating weight, which utilizes deep neural networks to extract weight-related features from a segmented mask obtained from a 2D image. Two image segmentation methods, fully and weakly supervised segmentation, were compared, with the first method achieving a more precise mask but requiring ground truth labeling. The experimental results showed a mean average error of 17.31 kg and mean absolute percentage error of 5.52% for fully supervised segmentation, and a mean average error of 35.91 kg and mean absolute percentage error of 10.1% for weakly supervised segmentation.
APPLIED SCIENCES-BASEL
(2023)
Article
Thermodynamics
Yavuz Selim Taspinar
Summary: Thermal energy in the infrared range can be detected and transformed into an image by thermal cameras, allowing non-contact detection of objects. This study proposed five different classification models and obtained new low-dimensional images by extracting features using HOG, LBP, SIFT, and GF methods. These images were classified using a CNN model called LW-CNN. The LW-CNN model achieved the highest classification accuracy in the evaluations, with accuracies of 98.58%, 95.56%, and 100% for the three datasets.
CASE STUDIES IN THERMAL ENGINEERING
(2023)
Article
Computer Science, Information Systems
Tim Oblak, Jaka Sircelj, Vitomir Struc, Peter Peer, Franc Solina, Ales Jaklic
Summary: The article addresses the challenges of reconstructing 3D space from visual data by focusing on recovering volumetric primitives from depth images. It introduces a novel solution for recovering superquadrics, a special type of parametric model capable of describing a wide array of 3D shapes using only a few parameters. Two learning strategies for training CNN predictors for superquadric parameters are developed and evaluated on a large dataset of synthetic and real-world depth images, showing favorable results compared to existing state-of-the-art techniques.
Article
Pediatrics
Anjum Shakya, Nisha Keshary Bhatta, Rupa Rajbhandari Singh, Shankar Prasad Yadav, Jitendra Thakur
Summary: This study aimed to develop gestational age specific percentile charts of birth weight, length, and head circumference for neonates. The percentile charts can be used as a reference for the local population and potentially establish a national reference curve for healthy neonates across different regions.
Article
Automation & Control Systems
Emmanuel Pintelas, Panagiotis Pintelas
Summary: This study combines unsupervised and supervised representation/feature learning methods to build a hybrid deep representation learning framework model applied on 3D Image inputs, and evaluates its performance on DeepFake and Pneumonia detection problems.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2022)
Article
Public, Environmental & Occupational Health
Abdolreza Marefat, Mahdieh Marefat, Javad Hassannataj Joloudari, Mohammad Ali Nematollahi, Reza Lashgari
Summary: COVID-19 is a novel virus that rapidly spreads and affects individuals' lives in various ways. Detecting the virus is crucial, and medical imaging such as CT and X-ray images are commonly used. However, the current procedures and high caseloads present challenges for medical practitioners. In this study, we propose a transformer-based method using Compact Convolutional Transformers (CCT) for automatically detecting COVID-19 from X-ray images. Our experiments demonstrate the effectiveness of the method with an accuracy of 99.22%, outperforming previous works.
FRONTIERS IN PUBLIC HEALTH
(2023)
Article
Computer Science, Interdisciplinary Applications
Hans Pinckaers, Wouter Bulten, Jeroen van der Laak, Geert Litjens
Summary: Prostate cancer is the most prevalent cancer among men in Western countries, and pathologists' evaluation is the gold standard for diagnosis. State-of-the-art convolutional neural networks are often patch-based and require detailed pixel-level annotations for effective training.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2021)
Article
Environmental Sciences
Mingzhe Feng, Xin Sun, Junyu Dong, Haoran Zhao
Summary: This paper proposes a network structure that uses dynamic receptive field and Gaussian pyramid pooling to address the issue of scale variation in remote sensing image segmentation. The network achieves better performance than other methods on large remote sensing image datasets.
Review
Environmental Sciences
D. P. Zielinski, C. Freiburger
Summary: Addressing the impact of dams and other water control structures on fish communities and aquatic ecosystems is a major concern for fisheries managers in the Laurentian Great Lakes. The current fish passage applications are vastly outnumbered by barriers to fish movement, highlighting the need for alternative management actions surrounding increased connectivity and invasive species control.
JOURNAL OF GREAT LAKES RESEARCH
(2021)
Article
Biology
Daniel P. Zielinski, Robert L. Mclaughlin, Thomas C. Pratt, R. Andrew Goodwin, Andrew M. Muir
Article
Fisheries
Sean A. Lewandoski, Peter Hrodey, Scott Miehls, Paul P. Piszczek, Daniel P. Zielinski
Summary: This study used computational fluid dynamics modeling and competitive risk analysis to develop predictive selective passage models. It found that sea lamprey's upstream passage probability decreased sharply as flow conditions became more turbulent, while declines in native fish white sucker were less substantial. Deploying a sea lamprey trap in fishways did not effectively reduce sea lamprey's upstream passage probability, but using bifurcated fishways could improve trapping effectiveness.
CANADIAN JOURNAL OF FISHERIES AND AQUATIC SCIENCES
(2021)
Article
Environmental Sciences
Reid G. Swanson, Erin L. McCann, Nicholas S. Johnson, Daniel P. Zielinski
Summary: Research suggests that medium-sized fish, mainly white sucker Catostomus commersonii and longnose sucker Catostomus catostomus, are more likely to enter the river at sunset and less likely at midnight in the Boardman River, a tributary of Lake Michigan. The entry rates of medium fish increase with rising river temperature and discharge, but decrease with higher lake levels. This understanding is crucial in developing fish passage solutions and management regulations for Great Lakes migratory fishes.
JOURNAL OF GREAT LAKES RESEARCH
(2021)
Article
Environmental Sciences
Daniel P. Zielinski, Scott Miehls, Sean Lewandoski
Summary: Barriers are effective for managing invasive species but limit the migration of native species. Optical sorting and the Archimedes screw show potential for selectively passing fish and capturing invasive species. A field-scale prototype of the fish lift successfully transported a large number of fish, and the passage of suckers increased with water temperature and attraction flow.
Article
Fisheries
Prathyush Nallamothu, Jonathan Gregory, Jordan Leh, Daniel P. Zielinski, Jesse L. Eickholt
Summary: This study investigated the leap characteristics of rainbow trout present in the Laurentian Great Lakes. Data collection and annotation were conducted using a custom web application, and key markers were labeled to calculate launch speed, launch angle, and length of the fish. The findings showed that the launch angle of steelhead aligns closely with the direction of water velocity.
Article
Fisheries
Daniel Patrick Zielinski, Peter W. Sorensen
Summary: By modifying spillway gate operation, adding deterrent systems, and removing carp, the upstream movement rates of invasive bigheaded carp can be reduced significantly. Implementing a combination of these measures can lead to a 98-99% reduction in carp movement, effectively containing their spread and improving blockage effectiveness.
Article
Fisheries
Jesse Eickholt, Dylan Kelly, Janine Bryan, Scott Miehls, Dan Zielinski
ICES JOURNAL OF MARINE SCIENCE
(2020)
