Hyperspectral Tissue Image Segmentation Using Semi-Supervised NMF and Hierarchical Clustering

Kumar, Neeraj; Uppala, Phanikrishna; Duddu, Karthik; Sreedhar, Hari; Varma, Vishal K.; Guzman, Grace; Sethi, Amit

doi:10.1109/tmi.2018.2883301

Cited by 36 publications

(18 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…11 Kumar et al applied semi-supervised NMF in spectral dimension reduction and hierarchical pixel clustering, the classi¯cation method they proposed can explain the discrepancies in the spectra of sub-cellular components of the same cell type. 12 These studies manifest that the spectral information e®ectively improves classi¯cation accuracy. Recently, with the rapid development of deep learning algorithms, the research and application of biomedical image processing have appeared.…”

Section: Introductionmentioning

confidence: 93%

Spatial-spectral identification of abnormal leukocytes based on microscopic hyperspectral imaging technology

Zhou

et al. 2020

J. Innov. Opt. Health Sci.

View full text Add to dashboard Cite

Screening and diagnosing of abnormal Leukocytes are crucial for the diagnosis of immune diseases and Acute Lymphoblastic Leukemia (ALL). As the deterioration of abnormal leukocytes is mainly due to the changes in the chromatin distribution, which signi¯cantly a®ects the absorption and re°ection of light, the spectral feature is proved to be important for leukocytes classi¯cation and identi¯cation. This paper proposes an accurate identi¯cation method for healthy and abnormal leukocytes based on microscopic hyperspectral imaging (HSI) technology which combines the spectral information. The segmentation of nucleus and cytoplasm is obtained by the morphological watershed algorithm. Then, the spectral features are extracted and combined with the spatial features. Based on this, the support vector machine (SVM) is applied for classi¯cation of¯ve types of leukocytes and abnormal leukocytes. Compared with di®erent classi¯cation methods, the proposed method utilizes spectral features which highlight the differences between healthy leukocytes and abnormal leukocytes, improving the accuracy in the classi¯cation and identi¯cation of leukocytes. This paper only selects one subtype of ALL for test, and the proposed method can be applied for detection of other leukemia in the future.

show abstract

Section: Introductionmentioning

confidence: 93%

Spatial-spectral identification of abnormal leukocytes based on microscopic hyperspectral imaging technology

Zhou

et al. 2020

J. Innov. Opt. Health Sci.

View full text Add to dashboard Cite

show abstract

“…At each intermediate stage, a new partition is created by linking the two most similar existing partitions. HAC has been applied to hyperspectral images [11], [12], [13], [14]; however, it has not been applied to spectral libraries for target identification.…”

Section: A Hierarchical Clusteringmentioning

confidence: 99%

Efficient Hyperspectral Target Detection and Identification With Large Spectral Libraries

Loughlin

Pieper

Manolakis

et al. 2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Numerous hyperspectral algorithms have been developed to detect both full and sub-pixel solid target materials. Target signatures are obtained from spectral libraries that contain both target and non-target materials. When the library is large and contains many potential targets, it is inefficient to run an individual detector for each material of interest. Additionally, such an approach produces numerous false alarms (i.e., multiple detections per pixel) due to spectral similarity among targets. In this paper, we present an efficient approach for detecting multiple targets within large spectral libraries while mitigating false alarms. We first group spectrally similar materials within the library into a hierarchy of clusters. From each cluster containing a target material, a single detector is obtained. Each detector represents multiple library spectra, so an identification step is needed for detected pixels. Detected pixels are modeled as a mixture between their local in-scene background and candidate library spectra. Candidates are chosen from adjacent library clusters. The candidate model providing the best fit is chosen to report. Use of local background spectra provides a physically meaningful mixing model that adapts to detected pixels. Clustering the library reduces the computational complexity of modeling detected pixels. We demonstrate detection and false alarm mitigation performance of our proposed algorithm with a real hyperspectral dataset.

show abstract

“…With the rapid development of intelligent technologies, automated knowledge discovery based on clustering becomes more and more important in these years. Although a large number of clustering algorithms have been successfully used in image segmentation and data classification [4], [5], it is still a challenging topic because it is difficult to achieve automatic clustering and to provide fine results for image segmentation. Image segmentation algorithms based on clustering have two advantages.…”

Section: Introductionmentioning

confidence: 99%

Automatic Fuzzy Clustering Framework for Image Segmentation

Lei

Liu

Jia

et al. 2020

IEEE Trans. Fuzzy Syst.

130

View full text Add to dashboard Cite

Clustering algorithms by minimizing an object function share a clear drawback that the number of clusters need to be set manually. Although density peak clustering is able to seek the number of clusters, it suffers from memory overflow when it is used for image segmentation because a moderate-size image usually includes a large number of pixels leading to a huge similarity matrix. To address the issue, here we proposed an automatic fuzzy clustering framework (AFCF) for image segmentation. The proposed framework has threefold contributions. Firstly, the idea of superpixel is used for the density peak (DP) algorithm, which efficiently reduces the size of the similarity matrix and thus improves the computational efficiency of the DP algorithm. Secondly, we employ a density balance algorithm to obtain a more robust decision-graph that helps the DP algorithm to achieve fully automatic clustering. Finally, a fuzzy c-means clustering based on prior entropy is used in the framework to improve image segmentation results. Because the spatial neighboring information of both the pixels and membership are considered, the final segmentation result is improved effectively. Experiments show that the proposed framework is not only able to achieve automatic image segmentation, it also provides better segmentation results than state-of-the-art algorithms.Index Terms-Fuzzy clustering, image segmentation, superpixel, density peak (DP) algorithm Tao Lei (M'17) received the Ph.D degree in Information and Communication Engineering from Northwestern

show abstract

Hyperspectral Tissue Image Segmentation Using Semi-Supervised NMF and Hierarchical Clustering

Cited by 36 publications

References 39 publications

Spatial-spectral identification of abnormal leukocytes based on microscopic hyperspectral imaging technology

Spatial-spectral identification of abnormal leukocytes based on microscopic hyperspectral imaging technology

Efficient Hyperspectral Target Detection and Identification With Large Spectral Libraries

Automatic Fuzzy Clustering Framework for Image Segmentation

Contact Info

Product

Resources

About