I tiakina i:
| Ngā kaituhi matua: | , , , , , |
|---|---|
| Hōputu: | Recurso digital |
| Reo: | Ingarihi |
| I whakaputaina: |
Zenodo
2026
|
| Ngā marau: | |
| Urunga tuihono: | https://doi.org/10.5281/zenodo.20394017 |
| Ngā Tūtohu: |
Tāpirihia he Tūtohu
Kāore He Tūtohu, Me noho koe te mea tuatahi ki te tūtohu i tēnei pūkete!
|
Rārangi ihirangi:
- <p class="MsoBodyText"><em>Brain<span> </span>tumors<span> </span>and<span> </span>cerebral<span> </span>blockages<span> </span>(which<span> </span>lead<span> </span>to<span> </span>strokes)<span> </span>are<span> </span>life-threatening<span> </span><span>conditions </span>requiring<span> </span>immediate<span> </span>and<span> </span>accurate<span> </span>diagnosis.<span> </span>Manual<span> </span>evaluation<span> </span>of<span> </span>MRI<span> </span>scans<span> </span>by<span> </span>radiologists<span> </span>is slow and subjective. This paper proposes an automated deep learning system using a Convolutional<span> </span>Neural<span> </span>Network<span> </span>(CNN)<span> </span>to<span> </span>classify brain<span> </span>MRI<span> </span>images<span> </span>into<span> </span><span>three </span>categories: Tumor, Blockage, or Normal.<span> </span>The system preprocesses images (grayscale, resizing, and normalization),<span> </span>extracts<span> </span>spatial<span> </span>features<span> </span>through<span> </span>convolutional<span> </span>and<span> </span>pooling<span> </span>layers,<span> </span>and<span> </span>outputs<span> </span>a prediction with confidence. The method reduces diagnostic time, eliminates manual feature engineering,<span> </span>and<span> </span>provides<span> </span>consistent<span> </span>results.<span> </span>Experimental<span> </span>evaluation<span> </span>on<span> </span>a<span> </span>mixed<span> </span>MRI<span> </span>dataset shows an expected accuracy of over 95%, demonstrating its potential as a clinical decision support <span>tool.</span></em></p> <p class="MsoBodyText"><em> </em></p>