Deep learning approach for hyperspectral image demosaicking, spectral correction and high-resolution RGB reconstruction
نویسندگان
چکیده
Hyperspectral imaging is one of the most promising techniques for intraoperative tissue characterisation. Snapshot mosaic cameras, which can capture hyperspectral data in a single exposure, have potential to make real-time system surgical decision-making possible. However, optimal exploitation captured requires solving an ill-posed demosaicking problem and applying additional spectral corrections. In this work, we propose supervised learning-based image algorithm snapshot images. Due lack publicly available medical images acquired with synthetic generation approach proposed simulate from existing datasets by high-resolution, but slow, devices. Image reconstruction achieved using convolutional neural networks super-resolution, followed correction sensor-specific calibration matrix. The results are evaluated both quantitatively qualitatively, showing clear improvements quality compared baseline method linear interpolation. Moreover, fast processing time 45 ms our obtain super-resolved RGB or oxygenation saturation maps per state-of-the-art camera demonstrates its seamless integration into applications.
منابع مشابه
An Effective Image Demosaicking Algorithm with Correlations among RGB Channels
In this paper, an effective image demosaicking algorithm, which is based on the correlation among the three primary colors, is proposed for mosaic image with Bayer color filter array (CFA). To reduce the distortion and improve the reconstruction quality, the proposed interpolation method makes full use of the brightness information and the edge information. We design several filters with size o...
متن کاملEvaluation of a hyperspectral image database for demosaicking purposes
We present a study on the the applicability of hyperspectral images to evaluate color filter array (CFA) design and the performance of demosaicking algorithms. The aim is to simulate a typical digital still camera processing pipe-line and to compare two different scenarios: evaluate the performance of demosaicking algorithms applied to raw camera RGB values before color rendering to sRGB, and e...
متن کاملSparse Spatio-spectral Representation for Hyperspectral Image Super-resolution
Existing hyperspectral imaging systems produce low spatial resolution images due to hardware constraints. We propose a sparse representation based approach for hyperspectral image super-resolution. The proposed approach first extracts distinct reflectance spectra of the scene from the available hyperspectral image. Then, the signal sparsity, non-negativity and the spatial structure in the scene...
متن کاملTraining-Based Spectral Reconstruction from a Single RGB Image
This paper focuses on a training-based method to reconstruct a scene’s spectral reflectance from a single RGB image captured by a camera with known spectral response. In particular, we explore a new strategy to use training images to model the mapping between cameraspecific RGB values and scene reflectance spectra. Our method is based on a radial basis function network that leverages RGB white-...
متن کاملHyperspectral image classification via contextual deep learning
Because the reliability of feature for every pixel determines the accuracy of classification, it is important to design a specialized feature mining algorithm for hyperspectral image classification. We propose a feature learning algorithm, contextual deep learning, which is extremely effective for hyperspectral image classification. On the one hand, the learning-based feature extraction algorit...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Computer methods in biomechanics and biomedical engineering. Imaging & visualization
سال: 2021
ISSN: ['2168-1171', '2168-1163']
DOI: https://doi.org/10.1080/21681163.2021.1997646