A5101570709- Brain Tumor Detection and Classification
- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- Topic Modeling
- Text and Document Classification Technologies
- Natural Language Processing Techniques
- Neurological Disease Mechanisms and Treatments
- Chinese history and philosophy
- COVID-19 diagnosis using AI
- Regulation of Appetite and Obesity
- Biochemical Analysis and Sensing Techniques
- Neuroinflammation and Neurodegeneration Mechanisms
- Cinema and Media Studies
- Neuroscience and Neuropharmacology Research
- Adipokines, Inflammation, and Metabolic Diseases
- Digital Media and Visual Art
Concordia University
2022-2024
Huanghe Science and Technology College
2014
First Affiliated Hospital of Harbin Medical University
2010
Harbin Medical University
2010
Recent pre-trained language models (PLMs) achieve promising results in existing abstractive summarization datasets. However, benchmarks overlap time with the standard pre-training corpora and finetuning Hence, strong performance of PLMs may rely on parametric knowledge that is memorized during fine-tuning. Moreover, by quickly become outdated, which affects generalization future data. In this work, we propose TempoSum, a novel benchmark contains data samples from 2010 to 2022, understand...
The influence of cilostazol on learning and memory, cyclin D1 expression in the cerebral cortex rats with chronic ischemia were investigated. A model was established using permanent bilateral common carotid artery occlusion method (2VO), memory capacity detected Morris water maze, changes apoptosis regulating gene tested by RT-PCR. Results maze indicated that significant extensions found escape latent period swimming path group (2VO group), results obviously superior compared to 2VO...
The research on CNN applications for medical image processing has been progressing rapidly. In the process of development, hurdles appear and are to be overcome. limitation in training samples is one them, restriction computation resources can another. this paper, we present a design approach application specific (ASCNN), allowing minimize computational complexity systems without lowering performance. This full-custom CNNs applications, such as brain tumor detection, so that each part...
Recent pre-trained language models (PLMs) achieve promising results in existing abstractive summarization datasets. However, benchmarks overlap time with the standard pre-training corpora and finetuning Hence, strong performance of PLMs may rely on parametric knowledge that is memorized during fine-tuning. Moreover, by quickly become outdated, which affects generalization future data. In this work, we propose TempoSum, a novel benchmark contains data samples from 2010 to 2022, understand...
The work presented in this paper is to propose a reliable high-quality system of Convolutional Neural Network (CNN) for brain tumor segmentation with low computation requirement. consists CNN the main processing segmentation, pre-CNN block data reduction and post-CNN refinement block. unique 7 convolution layers involving only 108 kernels 20308 trainable parameters. It custom-designed, following proposed paradigm ASCNN (application specific CNN), perform mono-modality cross-modality feature...
we can read film montage anywhere in great Chinese poetry.We also find the charm of poetry montage.The reason is that both are very common language structure and figures speech.Whether narrative, expressing emotions thought or metaphor, contrast symbol compression, amplification, crisscross stream-of-consciousness so on, all these artistic thoughts.Their forms thinking same similar.That an amazing realm.It worth further studying.It instructive for us to learn creation films TV plays.Let's...
In this paper, we propose a system to detect brain tumor in 3D MRI scans of Flair modality. It performs 2 functions: (a) predicting gray-level and locational distributions the pixels regions (b) generating mask pixel-wise precision. To facilitate data analysis processing, introduced 2D histogram presentation that comprehends distribution pixel-location object. proposed system, particular histograms, which tumor-related feature get concentrated, are established by exploiting left-right...