Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Mar 11, 2026

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

49.6K

Image Preprocessing and Optimizer Sensitivity: Implications for Convolutional Neural Networks in Diagnosing Brain

Manikanta Manohar Pattisapu1, Suraj Aravind B2, V Venkateswara Rao P3

  • 1Gandhi Institute of Technology and Management (GITAM) Deemed to be University; mpattisa@gitam.edu.

Journal of Visualized Experiments : Jove
|March 9, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same journal

Structure-Guided Design and Development of Novel Cyclophilin A Inhibitors and Ganoderiol-F Derivatives: An In-Silico Approach.

Journal of visualized experiments : JoVE·2026
Same journal

A 3D Bone Culture Platform Using Human Osteocytes and Decellularized Extracellular Matrix for Modeling Musculoskeletal Diseases.

Journal of visualized experiments : JoVE·2026
Same journal

Robotic Arm Assisted Slow Electrode Insertion and Simultaneous Measurement of Cochlear Microphonics in Cochlear Implantation.

Journal of visualized experiments : JoVE·2026
Same journal

Integrative Transcriptome-Wide Association Study and Mendelian Randomization Identify LYNX1 and MS4A14 as Therapeutic Targets for Osteomyelitis.

Journal of visualized experiments : JoVE·2026
Same journal

Synergistic Effect of Hypertension and Smoking on Ischemic Stroke Risk: A Case-Control Study With Additive and Multiplicative Interaction Analysis.

Journal of visualized experiments : JoVE·2026
Same journal

Research Progress on DNA Damage Repair Efficiency Assessment Strategies.

Journal of visualized experiments : JoVE·2026

Simple preprocessing for brain tumor classification MRI images improves CNN performance and optimizer convergence. Baseline preprocessing achieved 99.53% accuracy, highlighting its importance for robust medical image analysis.

Area of Science:

  • Medical Image Analysis
  • Machine Learning in Healthcare

Background:

  • Brain tumor classification via MRI is challenging due to image variability.
  • The impact of image preprocessing on optimizer behavior and CNN performance is understudied.

Purpose of the Study:

  • To investigate the effect of different MRI preprocessing pipelines on optimizer performance and CNN accuracy for brain tumor classification.
  • To compare baseline resizing versus traditional preprocessing methods (grayscale, blur, morphological filtering).

Main Methods:

  • Utilized a Kaggle MRI dataset for brain tumor classification.
  • Implemented two preprocessing pipelines: baseline (resizing) and traditional (grayscale, blur, filtering).
  • Evaluated three optimizers (Adam, RMSProp, SGD) with a fixed CNN architecture, using five-fold cross-validation.

Related Experiment Videos

Last Updated: Mar 11, 2026

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

49.6K

Main Results:

  • Baseline preprocessing consistently resulted in higher accuracy and more stable convergence across all tested optimizers.
  • RMSProp and SGD achieved the highest mean accuracy of 99.53% with baseline preprocessing.
  • The study quantified the impact of preprocessing choices on model performance metrics (accuracy, precision, recall, F1-score).

Conclusions:

  • Simple, baseline image preprocessing is more effective than traditional methods for brain tumor MRI classification.
  • Preprocessing strategies significantly influence optimizer convergence and CNN generalization.
  • Developing preprocessing-aware training is crucial for enhancing robustness and interpretability in medical imaging AI.