Get Started
Home
Topics
Search
Library
Research questionCan preprocessing defenses detect adversarial attacks in depthwise-separable edge vision CNNs when they cannot restore predictions?Preprocessing defenses are often assumed to transfer across model architectures, but depthwise-separable CNNs may respond differently to adversarial perturbations than residual or Inception-style networks. Their failure to recover predictions may still produce measurable differences between clean and adversarial inputs, while image-quality scores may not reflect defensive value.
AI
Computer Vision
Evaluation & Benchmarks
Image & Video Processing
Machine Learning
Research Paper
Small / On-device Models
Technology
Latest papersRecent research connected to this question, newest first.Preprocessing Failure and Adversarial Detection in Depthwise-Separable Edge Vision SystemsThe evidence covers six preprocessing defenses, two depthwise-separable architectures, one residual architecture, and a single attack family. It reports poor recovery for the depthwise-separable models, partial recovery for the residual model, and an observable clean-versus-adversarial output divergence; broader architectural and attack coverage is not established.research paper · Sep 3, 2026
Related questions
How can deployed vision models resist diverse backdoor triggers without model internals or auxiliary data?How can adversarial perturbations transfer to unseen semantic-segmentation models while accounting for dense spatial and class-wise structure?How can histopathology models withstand adversarial perturbations without erasing diagnostic tissue structures?Can small targeted grayscale patches force chosen semantics in infrared vision-language models across tasks?