Deep learning for liver tumor diagnosis part I: development of a convolutional neural network classifier for multi-phasic MRI

ABSTRACTIn this paper, deep learning is coupled with explainable artificial intelligence techniques for the discovery of representative genomic sequences in SARS-CoV-2. A convolutional neural network classifier is first trained on 553 sequences from available repositories, separating the genome of different virus strains from the Coronavirus family with considerable accuracy. The network’s behavior is then analyzed, to discover sequences used by the model to identify SARS-CoV-2, ultimately uncovering sequences exclusive to it. The discovered sequences are first validated on samples from other repositories, and proven able to separate SARS-CoV-2 from different virus strains with near-perfect accuracy. Next, one of the sequences is selected to generate a primer set, and tested against other state-of-the-art primer sets on existing datasets, obtaining competitive results. Finally, the primer is synthesized and tested on patient samples (n=6 previously tested positive), delivering a sensibility similar to routine diagnostic methods, and 100% specificity. In this paper, deep learning is coupled with explainable artificial intelligence techniques for the discovery of representative genomic sequences in SARS-CoV-2. A convolutional neural network classifier is first trained on 553 sequences from NGDC, separating the genome of different virus strains from the Coronavirus family with accuracy 98.73%. The network’s behavior is then analyzed, to discover sequences used by the model to identify SARS-CoV-2, ultimately uncovering sequences exclusive to it. The discovered sequences are validated on samples from NCBI and GISAID, and proven able to separate SARS-CoV-2 from different virus strains with near-perfect accuracy. Next, one of the sequences is selected to generate a primer set, and tested against other state-of-the-art primer sets, obtaining competitive results. Finally, the primer is synthesized and tested on patient samples (n=6 previously tested positive), delivering a sensibility similar to routine diagnostic methods, and 100% specificity. The proposed methodology has a substantial added value over existing methods, as it is able to both identify promising primer sets for a virus from a limited amount of data, and deliver effective results in a minimal amount of time. Considering the possibility of future pandemics, these characteristics are invaluable to promptly create specific detection methods for diagnostics.

Download Full-text

Doctor's Dilemma: Evaluating an Explainable Subtractive Spatial Lightweight Convolutional Neural Network for Brain Tumor Diagnosis

ACM Transactions on Multimedia Computing Communications and Applications ◽

10.1145/3457187 ◽

2021 ◽

Vol 17 (3s) ◽

pp. 1-26

Author(s):

Ambeshwar Kumar ◽

Ramachandran Manikandan ◽

Utku Kose ◽

Deepak Gupta ◽

Suresh C. Satapathy

Keyword(s):

Neural Network ◽

Deep Learning ◽

Brain Tumor ◽

Convolutional Neural Network ◽

Classification Accuracy ◽

Medical Image ◽

Tumor Classification ◽

Tumor Diagnosis ◽

Trust Level ◽

Brain Tumor Diagnosis

In Medicine Deep Learning has become an essential tool to achieve outstanding diagnosis on image data. However, one critical problem is that Deep Learning comes with complicated, black-box models so it is not possible to analyze their trust level directly. So, Explainable Artificial Intelligence (XAI) methods are used to build additional interfaces for explaining how the model has reached the outputs by moving from the input data. Of course, that's again another competitive problem to analyze if such methods are successful according to the human view. So, this paper comes with two important research efforts: (1) to build an explainable deep learning model targeting medical image analysis, and (2) to evaluate the trust level of this model via several evaluation works including human contribution. The target problem was selected as the brain tumor classification, which is a remarkable, competitive medical image-based problem for Deep Learning. In the study, MR-based pre-processed brain images were received by the Subtractive Spatial Lightweight Convolutional Neural Network (SSLW-CNN) model, which includes additional operators to reduce the complexity of classification. In order to ensure the explainable background, the model also included Class Activation Mapping (CAM). It is important to evaluate the trust level of a successful model. So, numerical success rates of the SSLW-CNN were evaluated based on the peak signal-to-noise ratio (PSNR), computational time, computational overhead, and brain tumor classification accuracy. The objective of the proposed SSLW-CNN model is to obtain faster and good tumor classification with lesser time. The results illustrate that the SSLW-CNN model provides better performance of PSNR which is enhanced by 8%, classification accuracy is improved by 33%, computation time is reduced by 19%, computation overhead is decreased by 23%, and classification time is minimized by 13%, as compared to state-of-the-art works. Because the model provided good numerical results, it was then evaluated in terms of XAI perspective by including doctor-model based evaluations such as feedback CAM visualizations, usability, expert surveys, comparisons of CAM with other XAI methods, and manual diagnosis comparison. The results show that the SSLW-CNN provides good performance on brain tumor diagnosis and ensures a trustworthy solution for the doctors.

Download Full-text

Convolutional Neural Network of Atomic Surface Structures to Predict Binding Energies for High-Throughput Screening of Catalysts

10.26434/chemrxiv.8150666.v1 ◽

2019 ◽

Author(s):

Seoin Back ◽

Junwoong Yoon ◽

Nianhan Tian ◽

Wen Zhong ◽

Kevin Tran ◽

...

Keyword(s):

Neural Network ◽

Deep Learning ◽

Convolutional Neural Network ◽

High Throughput ◽

High Throughput Screening ◽

Binding Energies ◽

Surface Structures ◽

Voronoi Polyhedra ◽

Atomic Surface

We present an application of deep-learning convolutional neural network of atomic surface structures using atomic and Voronoi polyhedra-based neighbor information to predict adsorbate binding energies for the application in catalysis.

Download Full-text

A Review for Investigation Studies That are Done for Improving Image ProcessingClassification Based on Convolutional Neural Network (CNN) That is Architectural of Deep Learning

International Congress on Human-Computer Interaction, Optimization and Robotic Applications Proceedings ◽

10.36287/setsci.4.5.007 ◽

2019 ◽

Author(s):

Mustafa Tüfekçi ◽

Fatih Karpat

Keyword(s):

Neural Network ◽

Deep Learning ◽

Convolutional Neural Network

Download Full-text

Effectiveness of transfer learning for enhancing tumor classification with a convolutional neural network on frozen sections

Scientific Reports ◽

10.1038/s41598-020-78129-0 ◽

2020 ◽

Vol 10 (1) ◽

Author(s):

Young-Gon Kim ◽

Sungchul Kim ◽

Cristina Eunbee Cho ◽

In Hye Song ◽

Hee Jin Lee ◽

...

Keyword(s):

Neural Network ◽

Deep Learning ◽

Convolutional Neural Network ◽

Transfer Learning ◽

Frozen Section ◽

Medical Center ◽

External Validation ◽

Model Performance ◽

Classification Model ◽

Training Dataset

AbstractFast and accurate confirmation of metastasis on the frozen tissue section of intraoperative sentinel lymph node biopsy is an essential tool for critical surgical decisions. However, accurate diagnosis by pathologists is difficult within the time limitations. Training a robust and accurate deep learning model is also difficult owing to the limited number of frozen datasets with high quality labels. To overcome these issues, we validated the effectiveness of transfer learning from CAMELYON16 to improve performance of the convolutional neural network (CNN)-based classification model on our frozen dataset (N = 297) from Asan Medical Center (AMC). Among the 297 whole slide images (WSIs), 157 and 40 WSIs were used to train deep learning models with different dataset ratios at 2, 4, 8, 20, 40, and 100%. The remaining, i.e., 100 WSIs, were used to validate model performance in terms of patch- and slide-level classification. An additional 228 WSIs from Seoul National University Bundang Hospital (SNUBH) were used as an external validation. Three initial weights, i.e., scratch-based (random initialization), ImageNet-based, and CAMELYON16-based models were used to validate their effectiveness in external validation. In the patch-level classification results on the AMC dataset, CAMELYON16-based models trained with a small dataset (up to 40%, i.e., 62 WSIs) showed a significantly higher area under the curve (AUC) of 0.929 than those of the scratch- and ImageNet-based models at 0.897 and 0.919, respectively, while CAMELYON16-based and ImageNet-based models trained with 100% of the training dataset showed comparable AUCs at 0.944 and 0.943, respectively. For the external validation, CAMELYON16-based models showed higher AUCs than those of the scratch- and ImageNet-based models. Model performance for slide feasibility of the transfer learning to enhance model performance was validated in the case of frozen section datasets with limited numbers.

Download Full-text

A Framework for Extractive Text Summarization based on Deep Learning Modified Neural Network Classifier

ACM Transactions on Asian and Low-Resource Language Information Processing ◽

10.1145/3392048 ◽

2020 ◽

Author(s):

BalaAnand Muthu ◽

Sivaparthipan CB ◽

Priyan Malarvizhi Kumar ◽

Seifedine Nimer Kadry ◽

Ching-Hsien Hsu ◽

...

Keyword(s):

Neural Network ◽

Deep Learning ◽

Text Summarization ◽

Neural Network Classifier

Download Full-text