Porting a PCA-based hyperspectral image dimensionality reduction algorithm for brain cancer detection on a manycore architecture

Lazcano, Raquel; Madroñal, Daniel; Salvador, Rubén; Desnos, Karol; Pelcat, Maxime; Guerra, Raúl; Fabelo, Himar; Ortega, Samuel; López, S.; Callicó, Gustavo M.; Juárez, E.; Sanz, C.

doi:10.1016/j.sysarc.2017.05.001

Cited by 38 publications

(25 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They also contain 2 MB of shared memory organized in 16 parallel banks of 128 KB each. This shared memory provides a bandwidth of 38.4 GB/s, providing a fair trade-off between power, area, latency, and bandwidth [21], although it has been demonstrated to be the main limitation in data-intensive applications [23]. Each cluster also gathers a DMA engine for handling the communications between the NoC and the shared memory.…”

Section: Preliminary Issuesmentioning

confidence: 99%

See 1 more Smart Citation

Implementation of the Principal Component Analysis onto High-Performance Computer Facilities for Hyperspectral Dimensionality Reduction: Results and Comparisons

et al. 2018

View full text Add to dashboard Cite

Dimensionality reduction represents a critical preprocessing step in order to increase the efficiency and the performance of many hyperspectral imaging algorithms. However, dimensionality reduction algorithms, such as the Principal Component Analysis (PCA), suffer from their computationally demanding nature, becoming advisable for their implementation onto high-performance computer architectures for applications under strict latency constraints. This work presents the implementation of the PCA algorithm onto two different high-performance devices, namely, an NVIDIA Graphics Processing Unit (GPU) and a Kalray manycore, uncovering a highly valuable set of tips and tricks in order to take full advantage of the inherent parallelism of these high-performance computing platforms, and hence, reducing the time that is required to process a given hyperspectral image. Moreover, the achieved results obtained with different hyperspectral images have been compared with the ones that were obtained with a field programmable gate array (FPGA)-based implementation of the PCA algorithm that has been recently published, providing, for the first time in the literature, a comprehensive analysis in order to highlight the pros and cons of each option.

show abstract

Section: Preliminary Issuesmentioning

confidence: 99%

“…This stage calculates the covariance matrix associated to the original image after being mean-centered. As exposed in previous works, this stage becomes the main bottleneck of the algorithm in platforms that are memory bounded, as the MPPA-256-N [23].…”

mentioning

confidence: 99%

Implementation of the Principal Component Analysis onto High-Performance Computer Facilities for Hyperspectral Dimensionality Reduction: Results and Comparisons

et al. 2018

View full text Add to dashboard Cite

show abstract

“…PCA is combined with EPF (PCA-EPFs) overcome the above issue and the results accuracy achieved very high with Support Vector Machine (SVM) classification [9]. In [10], for the targeted brain surgery application of HS data PCA parallelization technique is evolved in minimizing the time for capturing the frame-per minute. The vector quantization and PCA technique are combined to de-correlate the process of spectral information to compress the HS image ineffective way [11].…”

Section: A Principal Component Analysis (Pca)mentioning

confidence: 99%

The Novel Gravitational Mass Weighted PCA Technique for Feature Extraction in Hyperspectral Data Classification

Nandhini¹,

Porkodi²

2019

IJEAT

View full text Add to dashboard Cite

Recent advancements in the imaging spectrometer collect both spatial and spectral information which creates a huge dimensionality. The heavy spectral information creates to build a classifier for discerning between the materials in the scene. The minimum number of training labels always in an exchange between the spectral information and the performance is called the Hughes effect. Also the redundant of spectral information and noisy data presents in the hyperspectral scene. The above issues are overcome using feature extraction and feature selection methods which play a major role in the reduction of dimensionality. This paper proposes the novel fusion Gravitational Mass Weighted Principal Component Analysis (GMWPCA) techniques for hyperspectral data dimensionality. Also, this paper presents the deep insight about the feature extraction techniques in hyperspectral data of both supervised and unsupervised learning methods and experimental analysis in AVIRIS Indian Pines hyperspectral dataset by employing PCA, Probability PCA, LDA, and proposed techniques. The 93.63 % high accuracy achieved by using a novel proposed method.

show abstract

“…Hyperspectral images, known as hypercubes, contain rich information on a wide range of spectra with a high spectral resolution [7], hence, dimensionality reduction, image processing, and machine learning techniques are applied to extract the useful information from the vast amounts of HSI data, and have made many of the advancements in cancer identification: (1) Dimensionality reduction techniques. The principal component analysis [8,9], tensor decompositions [10], and T-distributed stochastic neighbor approach [11,12], were to reduce the dimensionality of features in hyperspectral images for compact expression; (2) Image processing techniques. Fourier coefficients [13], normalized difference nuclear index [14], sparse representation [15], box-plot and the watershed method [16], superpixel method [9], markov random fields [17,18], and morphological method [19], were used for hyperspectral image processing and quantification analysis; (3) Machine learning techniques.…”

Section: Introductionmentioning

confidence: 99%

Adaptive deep learning for head and neck cancer detection using hyperspectral imaging

Wang

et al. 2019

Vis. Comput. Ind. Biomed. Art

View full text Add to dashboard Cite

It can be challenging to detect tumor margins during surgery for complete resection. The purpose of this work is to develop a novel learning method that learns the difference between the tumor and benign tissue adaptively for cancer detection on hyperspectral images in an animal model. Specifically, an auto-encoder network is trained based on the wavelength bands on hyperspectral images to extract the deep information to create a pixel-wise prediction of cancerous and benign pixel. According to the output hypothesis of each pixel, the misclassified pixels would be reclassified in the right prediction direction based on their adaptive weights. The auto-encoder network is again trained based on these updated pixels. The learner can adaptively improve the ability to identify the cancer and benign tissue by focusing on the misclassified pixels, and thus can improve the detection performance. The adaptive deep learning method highlighting the tumor region proved to be accurate in detecting the tumor boundary on hyperspectral images and achieved a sensitivity of 92.32% and a specificity of 91.31% in our animal experiments. This adaptive learning method on hyperspectral imaging has the potential to provide a noninvasive tool for tumor detection, especially, for the tumor whose margin is indistinct and irregular.

show abstract

Porting a PCA-based hyperspectral image dimensionality reduction algorithm for brain cancer detection on a manycore architecture

Cited by 38 publications

References 21 publications

Implementation of the Principal Component Analysis onto High-Performance Computer Facilities for Hyperspectral Dimensionality Reduction: Results and Comparisons

Implementation of the Principal Component Analysis onto High-Performance Computer Facilities for Hyperspectral Dimensionality Reduction: Results and Comparisons

The Novel Gravitational Mass Weighted PCA Technique for Feature Extraction in Hyperspectral Data Classification

Adaptive deep learning for head and neck cancer detection using hyperspectral imaging

Contact Info

Product

Resources

About