Hierarchical mix-pooling and its applications to biomedical image classification

Siyamalan Manivannan; Ruixuan Wang; Emanuele Trucco

doi:10.1109/ISBI.2016.7493326

Hierarchical mix-pooling and its applications to biomedical image classification

Siyamalan Manivannan (Lead / Corresponding author), Ruixuan Wang, Emanuele Trucco

Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Citation (Scopus)

267 Downloads (Pure)

Abstract

This paper introduces Hierarchical Mix-pooling (HMP), a translation-invariant image representation improving the discriminative power of pooling representations by capturing intermediate-size structure information in images. HMP consists of two levels, one traditional pooling (e.g., sum pooling) applied to intermediate-size regions to collect the statistics of local features, and one different pooling (e.g., max pooling) collecting statistics of the previously region-based pooled results. Classification experiments show that HMP considerably improves accuracies with much smaller sizes of dictionaries compared to traditional pooling. The superior performance of HMP is confirmed by experiments with different local features and classifiers on two public biomedical datasets (ICPR HEp-2 cells and IRMA radiology).

Original language	English
Title of host publication	2016 IEEE 13th International Symposium on Biomedical Imaging
Subtitle of host publication	From Nano to Macro, ISBI 2016 - Proceedings
Publisher	IEEE
Pages	541-544
Number of pages	4
Volume	2016-June
ISBN (Electronic)	9781479923496
ISBN (Print)	9781479923502
DOIs	https://doi.org/10.1109/ISBI.2016.7493326
Publication status	Published - 16 Jun 2016
Event	2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Prague, Czech Republic Duration: 13 Apr 2016 → 16 Apr 2016

Conference

Conference	2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016
Country/Territory	Czech Republic
City	Prague
Period	13/04/16 → 16/04/16

Keywords

Dictionaries
Feature extraction
Biomedical imaging
Image coding
Encoding
Image representation
Support vector machines

ASJC Scopus subject areas

Biomedical Engineering
Radiology Nuclear Medicine and imaging

Access to Document

10.1109/ISBI.2016.7493326

Author Accepted ManuscriptAccepted author manuscript, 311 KB

Cite this

@inproceedings{39488f2a316740a2ad5c368c018ff7a8,

title = "Hierarchical mix-pooling and its applications to biomedical image classification",

abstract = "This paper introduces Hierarchical Mix-pooling (HMP), a translation-invariant image representation improving the discriminative power of pooling representations by capturing intermediate-size structure information in images. HMP consists of two levels, one traditional pooling (e.g., sum pooling) applied to intermediate-size regions to collect the statistics of local features, and one different pooling (e.g., max pooling) collecting statistics of the previously region-based pooled results. Classification experiments show that HMP considerably improves accuracies with much smaller sizes of dictionaries compared to traditional pooling. The superior performance of HMP is confirmed by experiments with different local features and classifiers on two public biomedical datasets (ICPR HEp-2 cells and IRMA radiology).",

keywords = "Dictionaries, Feature extraction, Biomedical imaging, Image coding, Encoding, Image representation, Support vector machines",

author = "Siyamalan Manivannan and Ruixuan Wang and Emanuele Trucco",

note = "This work was partially funded by 2011-2016 EU FP7 ERC project “CODIR: colonic disease investigation by robotic hydrocolonoscopy”, collaborative between the Universities of Dundee (PI Prof Sir A Cuschieri) and Leeds (PI Prof A Neville).; 2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 ; Conference date: 13-04-2016 Through 16-04-2016",

year = "2016",

month = jun,

day = "16",

doi = "10.1109/ISBI.2016.7493326",

language = "English",

isbn = "9781479923502",

volume = "2016-June",

pages = "541--544",

booktitle = "2016 IEEE 13th International Symposium on Biomedical Imaging",

publisher = "IEEE",

}

Manivannan, S, Wang, R & Trucco, E 2016, Hierarchical mix-pooling and its applications to biomedical image classification. in 2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings. vol. 2016-June, 7493326, IEEE, pp. 541-544, 2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016, Prague, Czech Republic, 13/04/16. https://doi.org/10.1109/ISBI.2016.7493326

Hierarchical mix-pooling and its applications to biomedical image classification. / Manivannan, Siyamalan (Lead / Corresponding author); Wang, Ruixuan; Trucco, Emanuele.
2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings. Vol. 2016-June IEEE, 2016. p. 541-544 7493326.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Hierarchical mix-pooling and its applications to biomedical image classification

AU - Manivannan, Siyamalan

AU - Wang, Ruixuan

AU - Trucco, Emanuele

N1 - This work was partially funded by 2011-2016 EU FP7 ERC project “CODIR: colonic disease investigation by robotic hydrocolonoscopy”, collaborative between the Universities of Dundee (PI Prof Sir A Cuschieri) and Leeds (PI Prof A Neville).

PY - 2016/6/16

Y1 - 2016/6/16

N2 - This paper introduces Hierarchical Mix-pooling (HMP), a translation-invariant image representation improving the discriminative power of pooling representations by capturing intermediate-size structure information in images. HMP consists of two levels, one traditional pooling (e.g., sum pooling) applied to intermediate-size regions to collect the statistics of local features, and one different pooling (e.g., max pooling) collecting statistics of the previously region-based pooled results. Classification experiments show that HMP considerably improves accuracies with much smaller sizes of dictionaries compared to traditional pooling. The superior performance of HMP is confirmed by experiments with different local features and classifiers on two public biomedical datasets (ICPR HEp-2 cells and IRMA radiology).

AB - This paper introduces Hierarchical Mix-pooling (HMP), a translation-invariant image representation improving the discriminative power of pooling representations by capturing intermediate-size structure information in images. HMP consists of two levels, one traditional pooling (e.g., sum pooling) applied to intermediate-size regions to collect the statistics of local features, and one different pooling (e.g., max pooling) collecting statistics of the previously region-based pooled results. Classification experiments show that HMP considerably improves accuracies with much smaller sizes of dictionaries compared to traditional pooling. The superior performance of HMP is confirmed by experiments with different local features and classifiers on two public biomedical datasets (ICPR HEp-2 cells and IRMA radiology).

KW - Dictionaries

KW - Feature extraction

KW - Biomedical imaging

KW - Image coding

KW - Encoding

KW - Image representation

KW - Support vector machines

UR - http://www.scopus.com/inward/record.url?scp=84978428088&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2016.7493326

DO - 10.1109/ISBI.2016.7493326

M3 - Conference contribution

AN - SCOPUS:84978428088

SN - 9781479923502

VL - 2016-June

SP - 541

EP - 544

BT - 2016 IEEE 13th International Symposium on Biomedical Imaging

PB - IEEE

T2 - 2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016

Y2 - 13 April 2016 through 16 April 2016

ER -

Hierarchical mix-pooling and its applications to biomedical image classification

Abstract

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this