INTERNATIONAL JOURNAL OF NOVEL RESEARCH AND DEVELOPMENT International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2456-4184 | Impact factor: 8.76 | ESTD Year: 2016
Scholarly open access journals, Peer-reviewed, and Refereed Journals, Impact factor 8.76 (Calculate by google scholar and Semantic Scholar | AI-Powered Research Tool) , Multidisciplinary, Monthly, Indexing in all major database & Metadata, Citation Generator, Digital Object Identifier(DOI)
Segmenting Brain Tumor images is crucial for computer-assisted diagnosis. The key to effective segmentation is for the model to be able to both see the overall picture and the minute details, or to learn image characteristics that contain a lot of contexts while maintaining high spatial resolutions. The most popular techniques, U-Net and its variations, extract and fuse multi-scale information in order to reach this aim. The fused features performance is nonetheless constrained by their tiny effective receptive fields and emphasis on local visual signals. In this work, we use a variety of machine learning techniques to forecast the survival rate. To conduct segmentation, we use a 3D UNet++ architecture and combine channel and spatial attention with the decoder network. To forecast the length of each patient's survival, we extract certain unique radiomic parameters based on the geometry, position, and shape of the segmented tumor and integrate them with clinical data. To demonstrate the impact of each attribute on the prediction of overall survival (OS), we also conduct comprehensive studies. According to the experimental findings, the most important factors to determine the OS are clinical characteristics like age and radionics properties like the histogram, location, and shape of the necrosis area.
we offer Segtran, a different segmentation framework built on transformers and UNet++, which even at high feature resolutions have an infinite effective receptive field. Segtran's central component is a new squeeze-and-expansion UNet++, in which an expansion block learns a variety of representations while a squeezed attention block regulates the self-attention of transformers. We also provide a brand-new positional encoding approach for transformers that imposes an image continuity inductive bias.
"BRAIN TUMOR DETECTION USING U-NET++ SEGMENTATION TECHNIQUE", International Journal of Novel Research and Development (www.ijnrd.org), ISSN:2456-4184, Vol.8, Issue 7, page no.c761-c770, July-2023, Available :http://www.ijnrd.org/papers/IJNRD2307275.pdf
Downloads:
000118753
ISSN:
2456-4184 | IMPACT FACTOR: 8.76 Calculated By Google Scholar| ESTD YEAR: 2016
An International Scholarly Open Access Journal, Peer-Reviewed, Refereed Journal Impact Factor 8.76 Calculate by Google Scholar and Semantic Scholar | AI-Powered Research Tool, Multidisciplinary, Monthly, Multilanguage Journal Indexing in All Major Database & Metadata, Citation Generator
Facebook Twitter Instagram LinkedIn