About Scary Filter

Scary Filter is a test harness which demonstrates the capability of ScaryNet in filtering scary images in web applications. In this demonstration, Scary Filter will blur an image suspected of containing scary elements. Scary Filter is written in pure HTML5 and Javascript to maximize portability across different platform and devices.

It could run on any HTML5 compliant web browser on desktop computers, internet kiosks, mobile devices, and even on smart tv.

What is ScaryNet?

ScaryNet is a compact Artificial Convolutional Neural Network trained to identify scary images. It is designed to be fast and light so that it could be executed on resource-constrained environment such as web browser, mobile device or embedded devices.

ScaryNet is trained using TensorFlow with Keras, and deployed with the help of TensorFlow.js library.

Why ScaryNet?

Common use-cases and user for ScaryNet


Mohammad Hafiz bin Ismail,

For any enquiries, please send an email to deeplearn@uitm.edu.my

Referencing this work

You can download a BibTex reference file or you may refer to this work as:

Ismail, M. H. (2022). ScaryNet Filter - a test website for ScaryNet Convolutional Neural Network. Retrieved <DATE>, <YEAR>, from: https://demo.mobilepit.com/ai/scarynet. doi:10.5281/zenodo.5862059