paper-with-me

홈 › Papers

The Level Weighted Structural Similarity Loss: A Step Away from the MSE

2019-04-30 · Yingjing Lu

The Mean Square Error (MSE) has shown its strength when applied in deep generative models such as Auto-Encoders to model reconstruction loss. However, in image domain especially, the limitation of MSE is obvious: it assumes pixel independence and ignores spatial relationships of samples. This contradicts most architectures of Auto-Encoders which use convolutional layers to extract spatial dependent features. We base on the structural similarity metric (SSIM) and propose a novel level weighted structural similarity (LWSSIM) loss for convolutional Auto-Encoders. Experiments on common datasets on various Auto-Encoder variants show that our loss is able to outperform the MSE loss and the Vanilla SSIM loss. We also provide reasons why our model is able to succeed in cases where the standard SSIM loss fails.

📄 PDF Abstract BibTeX arXiv:1904.13362

Code (0)

등록된 구현이 없습니다.

Tasks

SSIM

Similar Papers 제목 키워드 기반

A Mutual-Structure Weighted Sub-Pixel Multimodal Optical Remote Sensing Image Matching Method

2025-08-14 · Tao Huang, Hongbo Pan, Nanxi Zhou, Siyuan Zou 외 arxiv

Sub-pixel matching of multimodal optical images is a critical step in combined application of multiple sensors. However structural noise and inconsistencies arising from variations in multimodal image responses usually l…

Image Matching

NeighCNN: A CNN based SAR Speckle Reduction using Feature preserving Loss Function

2021-08-26 · Praveen Ravirathinam, Darshan Agrawal, J. Jennifer Ranjani

Coherent imaging systems like synthetic aperture radar are susceptible to multiplicative noise that makes applications like automatic target recognition challenging. In this paper, NeighCNN, a deep learning-based speckle…

Structural-Aware Sentence Similarity with Recursive Optimal Transport

2020-01-28 · Zihao Wang, Yong Zhang, Hao Wu

Measuring sentence similarity is a classic topic in natural language processing. Light-weighted similarities are still of particular practical significance even when deep learning models have succeeded in many other task…

SentenceSentence SimilaritySTS

A Loss Function for Generative Neural Networks Based on Watson's Perceptual Model

2020-06-26 · Steffen Czolbe, Oswin Krause, Ingemar Cox, Christian Igel

To train Variational Autoencoders (VAEs) to generate realistic imagery requires a loss function that reflects human perception of image similarity. We propose such a loss function based on Watson's perceptual model, whic…

Translation

A Loss Function for Generative Neural Networks Based on Watson’s Perceptual Model

2020-12-01 · NeurIPS 2020 12 · Steffen Czolbe, Oswin Krause, Ingemar Cox, Christian Igel

To train Variational Autoencoders (VAEs) to generate realistic imagery requires a loss function that reflects human perception of image similarity. We propose such a loss function based on Watson's perceptual model, whi…

Translation