Browsing by Author "Tschannerl, Julius"

Now showing 1 - 3 of 3
  • Thumbnail Image
    ItemOpen Access
    Hyperspectral image reconstruction using multi-colour and time-multiplexed LED illumination
    (Elsevier, 2019-05-06) Tschannerl, Julius; Ren, Jinchang; Zhao, Huimin; Kao, Fu-Jen; Marshall, Stephen; Yuen, Peter W. T.
    The rapidly rising industrial interest in hyperspectral imaging (HSI) has generated an increased demand for cost effective HSI devices. We are demonstrating a mobile and low-cost multispectral imaging system, enabled by time-multiplexed RGB Light Emitting Diodes (LED) illumination, which operates at video framerate. Critically, a deep Multi-Layer Perceptron (MLP) with HSI prior in the spectral range of 400–950 nm is trained to reconstruct HSI data. We incorporate regularisation and dropout to compensate for overfitting in the largely ill-posed problem of reconstructing the HSI data. The MLP is characterised by a relatively simple design and low computational burden. Experimental results on 51 objects of various references and naturally occurring materials show the effectiveness of this approach in terms of reconstruction error and classification accuracy. We were also able to show that utilising additional colour channels to the three R, G and B channels adds increased value to the reconstruction.
  • Thumbnail Image
    ItemOpen Access
    MIMR-DGSA: unsupervised hyperspectral band selection based on information theory and a modified discrete gravitational search algorithm
    (Elsevier, 2019-02-15) Tschannerl, Julius; Ren, Jinchang; Yuen, Peter W. T.; Sun, Genyun; Zhao, Huimin; Yang, Zhijing; Wang, Zheng; Marshall, Stephen
    Band selection plays an important role in hyperspectral data analysis as it can improve the performance of data analysis without losing information about the constitution of the underlying data. We propose a MIMR-DGSA algorithm for band selection by following the Maximum-Information-Minimum-Redundancy (MIMR) criterion that maximises the information carried by individual features of a subset and minimises redundant information between them. Subsets are generated with a modified Discrete Gravitational Search Algorithm (DGSA) where we definine a neighbourhood concept for feature subsets. A fast algorithm for pairwise mutual information calculation that incorporates variable bandwidths of hyperspectral bands called VarBWFastMI is also developed. Classification results on three hyperspectral remote sensing datasets show that the proposed MIMR-DGSA performs similar to the original MIMR with Clonal Selection Algorithm (CSA) but is computationally more efficient and easier to handle as it has fewer parameters for tuning.
  • Thumbnail Image
    ItemOpen Access
    Novel Gumbel-Softmax trick enabled concrete autoencoder with entropy constraints for unsupervised hyperspectral band selection
    (IEEE, 2021-06-04) Sun, He; Ren, Jinchang; Zhao, Huimin; Yuen, Peter W. T.; Tschannerl, Julius
    As an important topic in hyperspectral image (HSI) analysis, band selection has attracted increasing attention in the last two decades for dimensionality reduction in HSI. With the great success of deep learning (DL)-based models recently, a robust unsupervised band selection (UBS) neural network is highly desired, particularly due to the lack of sufficient ground truth information to train the DL networks. Existing DL models for band selection either depend on the class label information or have unstable results via ranking the learned weights. To tackle these challenging issues, in this article, we propose a Gumbel-Softmax (GS) trick enabled concrete autoencoder-based UBS framework (CAE-UBS) for HSI, in which the learning process is featured by the introduced concrete random variables and the reconstruction loss. By searching from the generated potential band selection candidates from the concrete encoder, the optimal band subset can be selected based on an information entropy (IE) criterion. The idea of the CAE-UBS is quite straightforward, which does not rely on any complicated strategies or metrics. The robust performance on four publicly available datasets has validated the superiority of our CAE-UBS framework in the classification of the HSIs.