Welcome to FHEON Documentation.

FHEON

FHEON is a configurable framework for developing privacy-preserving convolutional neural networks (CNNs) under homomorphic encryption (HE). The framework is built on the Cheon–Kim–Kim–Song (CKKS) scheme, an approximate HE construction that supports efficient arithmetic over real-valued data. FHEON adopts the Residue Number System (RNS) variant of CKKS as implemented in OpenFHE, enabling scalable, highly efficient encrypted computation with precise control over numerical accuracy. This design allows floating-point–like operations to be evaluated directly on ciphertexts while minimizing performance overhead.

Leveraging this architecture, FHEON executes complete inference pipelines entirely in the encrypted domain, ensuring that sensitive inputs are never revealed during computation. By emphasizing configurability, efficiency, and end-to-end security, FHEON enables practical deployment of encrypted neural network inference in privacy-critical environments.

Key features of FHEON include:

Optimized HE CNN Layers: Provide users with multiple variants of secure convolution, average pooling, ReLU, and fully connected layers.
Configurable Architecture: All CNN Layers can be customized with standard paramters such input/output channels, kernel size, stride, and padding.
Versatile Evaluation: Provides multiple examples from multiple architectures including VGG-11, VGG-16, ResNet-20, and ResNet-34.

FHEON provides a flexible and efficient platform for researchers and developers to build HE-friendly neural networks with ease.

Citation

If you use FHEON in your work, please cite the following paper:

@misc{njungle2025fheonconfigurableframeworkdeveloping,
   title={FHEON: A Configurable Framework for Developing Privacy-Preserving Neural Networks Using Homomorphic Encryption},
   author={Nges Brian Njungle and Eric Jahns and Michel A. Kinsy},
   year={2025},
   eprint={2510.03996},
   archivePrefix={arXiv},
   primaryClass={cs.CR},
   url={https://arxiv.org/abs/2510.03996}

}

Link: Read the paper

CONTENT