Deep Demosaicing for Edge Implementation

Most digital cameras use sensors coated with a Color Filter Array (CFA) to capture channel components at every pixel location, resulting in a mosaic image that does not contain pixel values in all channels. Current research on reconstructing these missing channels, also known as demosaicing, introduces many artifacts, such as zipper effect and false color. Many deep learning demosaicing techniques outperform other classical techniques in reducing the impact of artifacts. However, most of these models tend to be over-parametrized. Consequently, edge implementation of the state-of-the-art deep learning-based demosaicing algorithms on low-end edge devices is a major challenge. We provide an exhaustive search of deep neural network architectures and obtain a pareto front of Color Peak Signal to Noise Ratio (CPSNR) as the performance criterion versus the number of parameters as the model complexity that beats the state-of-the-art. Architectures on the pareto front can then be used to choose the best architecture for a variety of resource constraints. Simple architecture search methods such as exhaustive search and grid search require some conditions of the loss function to converge to the optimum. We clarify these conditions in a brief theoretical study.


page 5

page 6


State of Compact Architecture Search For Deep Neural Networks

The design of compact deep neural networks is a crucial task to enable w...

POPNASv3: a Pareto-Optimal Neural Architecture Search Solution for Image and Time Series Classification

The automated machine learning (AutoML) field has become increasingly re...

BioNetExplorer: Architecture-Space Exploration of Bio-Signal Processing Deep Neural Networks for Wearables

In this work, we propose the BioNetExplorer framework to systematically ...

Resource Constrained Neural Network Architecture Search

The design of neural network architectures is frequently either based on...

Color Learning for Image Compression

Deep learning based image compression has gained a lot of momentum in re...

Monochrome and Color Polarization Demosaicking Using Edge-Aware Residual Interpolation

A division-of-focal-plane or microgrid image polarimeter enables us to a...

Grassroots Operator Search for Model Edge Adaptation

Hardware-aware Neural Architecture Search (HW-NAS) is increasingly being...

Please sign up or login with your details

Forgot password? Click here to reset