MisConv: Convolutional Neural Networks for Missing Data
This addresses a fundamental challenge in practical applications like image inpainting and autonomous systems, offering an efficient solution for handling missing data in CNNs.
The paper tackles the problem of processing missing data in convolutional neural networks (CNNs) by introducing MisConv, a mechanism that adapts CNNs to handle incomplete images using a Mixture of Factor Analyzers model, achieving superior or comparable performance to state-of-the-art methods in various image processing tasks.
Processing of missing data by modern neural networks, such as CNNs, remains a fundamental, yet unsolved challenge, which naturally arises in many practical applications, like image inpainting or autonomous vehicles and robots. While imputation-based techniques are still one of the most popular solutions, they frequently introduce unreliable information to the data and do not take into account the uncertainty of estimation, which may be destructive for a machine learning model. In this paper, we present MisConv, a general mechanism, for adapting various CNN architectures to process incomplete images. By modeling the distribution of missing values by the Mixture of Factor Analyzers, we cover the spectrum of possible replacements and find an analytical formula for the expected value of convolution operator applied to the incomplete image. The whole framework is realized by matrix operations, which makes MisConv extremely efficient in practice. Experiments performed on various image processing tasks demonstrate that MisConv achieves superior or comparable performance to the state-of-the-art methods.