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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Gastrointestinal image classification with GIDNet CNN model and non-linear Tansh activation function.

Computers in biology and medicine·2026
Same author

Evidence of spin-phonon-charge coupling and spin-orbit excitations in quasi-one-dimensional Ising spin chain system<i>α</i>-CoV<sub>2</sub>O<sub>6</sub>.

Journal of physics. Condensed matter : an Institute of Physics journal·2026
Same author

A hybrid CNN-transformer model with adaptive activation function for potato leaf disease classification.

Scientific reports·2026
Same author

Flatbands, Landau levels, Hall conductivity and Hofstadter butterfly in large-angle twisted bilayer graphene: a high pressure study.

Journal of physics. Condensed matter : an Institute of Physics journal·2025
Same author

Catalytic Asymmetric Total Synthesis of Naturally Occurring <i>Amaryllidaceae</i> Alkaloid, (+)-11-Hydroxyvittatine.

Organic letters·2025
Same author

Biomimetic total synthesis of the reported structure of (+)-selaginedorffone B.

Chemical science·2024

Related Experiment Video

Updated: Jul 12, 2025

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
14:08

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

42.7K

A novel Parametric Flatten-p Mish activation function based deep CNN model for brain tumor classification.

Ayan Mondal1, Vimal K Shrivastava1

  • 1School of Electronics Engineering, Kalinga Institute of Industrial Technology (KIIT), Bhubaneswar, India.

Computers in Biology and Medicine
|October 20, 2023
PubMed
Summary
This summary is machine-generated.

A novel Parametric Flatten-p Mish (PFpM) activation function enhances brain tumor classification accuracy in deep learning models. This new function improves upon existing methods, offering better performance in identifying brain tumors from medical images.

Keywords:
Activation functionBMRI-NetBrain tumor classificationCNNImage classificationMRIPFpM

More Related Videos

Visualization of Cortical Modules in Flattened Mammalian Cortices
08:49

Visualization of Cortical Modules in Flattened Mammalian Cortices

Published on: January 22, 2018

12.9K
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

48.1K

Related Experiment Videos

Last Updated: Jul 12, 2025

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
14:08

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

42.7K
Visualization of Cortical Modules in Flattened Mammalian Cortices
08:49

Visualization of Cortical Modules in Flattened Mammalian Cortices

Published on: January 22, 2018

12.9K
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

48.1K

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Computational Neuroscience

Background:

  • Brain tumors represent a significant global health challenge, being among the deadliest cancers.
  • Convolutional Neural Networks (CNNs) have shown great promise in medical image analysis.
  • Activation functions are critical components of CNNs, influencing model performance and learning capabilities.

Purpose of the Study:

  • To introduce a novel parametric activation function, Parametric Flatten-p Mish (PFpM), for improved brain tumor classification using CNNs.
  • To address limitations of existing activation functions, such as neuron death and bias shift.
  • To enhance the accuracy and flexibility of deep learning models in analyzing complex medical imaging data.

Main Methods:

  • Developed a CNN-based model, BMRI-Net, incorporating the proposed PFpM activation function.
  • Validated the model on two distinct brain tumor datasets: Figshare and Br35H.
  • Compared BMRI-Net's performance against state-of-the-art CNN models (DenseNet201, InceptionV3, MobileNetV2, ResNet50, VGG19) and various activation functions (ReLU, Leaky ReLU, GELU, Swish, Mish).
  • Conducted experiments using both record-wise and subject-wise data splits, employing hold-out and 5-fold cross-validation techniques.

Main Results:

  • BMRI-Net with PFpM achieved high overall accuracy, reaching up to 99.57% on the Figshare dataset (record-wise, hold-out validation).
  • Subject-wise validation on the Figshare dataset yielded an accuracy of 97.91% (hold-out validation).
  • On the Br35H dataset, the model attained 99% overall accuracy (record-wise, hold-out validation).
  • PFpM demonstrated superior performance compared to traditional activation functions across various experimental setups.

Conclusions:

  • The proposed PFpM activation function significantly improves the performance of CNNs for brain tumor classification.
  • BMRI-Net, powered by PFpM, offers a robust and accurate method for analyzing brain tumor images.
  • These findings suggest the potential of PFpM as a valuable tool for enhancing clinical diagnosis of brain tumors.