Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

CLASSIFICATION OF CORALS IN REFLECTANCE AND FLUORESCENCE IMAGES USING CONVOLUTIONAL NEURAL NETWORK REPRESENTATIONS

Abstract: 

Coral species, with complex morphology and ambiguous boundaries, pose a great challenge for automated classification. CNN activations, which are extracted from fully connected layers of deep networks (FC features), have been successfully used as powerful universal representations in many visual tasks. In this paper, we investigate the transferability and combined performance of FC features and CONV features (extracted
from convolutional layers) in the coral classification of two image modalities (reflectance and fluorescence), using a typical deep network (e.g. VGGNet). We exploit vector of locally aggregated descriptors (VLAD) encoding and principal component analysis (PCA) to compress dense CONV features into a compact representation. Experimental results demonstrate that encoded CONV3 features achieve superior performances on reflectance and fluorescence coral images, compared to FC features. The combination of these two features further improves the overall accuracy and achieves state-of-the-art performance on the challenging EFC dataset.

up
0 users have voted:

Paper Details

Authors:
Mohammed Bennamoun, Senjian An, Ferdous Sohel, Farid Boussaid
Submitted On:
20 April 2018 - 8:59am
Short Link:
Type:
Poster
Event:

Document Files

ICASSP2018 Poster.pdf

(52 downloads)

Subscribe

[1] Mohammed Bennamoun, Senjian An, Ferdous Sohel, Farid Boussaid, "CLASSIFICATION OF CORALS IN REFLECTANCE AND FLUORESCENCE IMAGES USING CONVOLUTIONAL NEURAL NETWORK REPRESENTATIONS", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3103. Accessed: Jul. 20, 2018.
@article{3103-18,
url = {http://sigport.org/3103},
author = {Mohammed Bennamoun; Senjian An; Ferdous Sohel; Farid Boussaid },
publisher = {IEEE SigPort},
title = {CLASSIFICATION OF CORALS IN REFLECTANCE AND FLUORESCENCE IMAGES USING CONVOLUTIONAL NEURAL NETWORK REPRESENTATIONS},
year = {2018} }
TY - EJOUR
T1 - CLASSIFICATION OF CORALS IN REFLECTANCE AND FLUORESCENCE IMAGES USING CONVOLUTIONAL NEURAL NETWORK REPRESENTATIONS
AU - Mohammed Bennamoun; Senjian An; Ferdous Sohel; Farid Boussaid
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3103
ER -
Mohammed Bennamoun, Senjian An, Ferdous Sohel, Farid Boussaid. (2018). CLASSIFICATION OF CORALS IN REFLECTANCE AND FLUORESCENCE IMAGES USING CONVOLUTIONAL NEURAL NETWORK REPRESENTATIONS. IEEE SigPort. http://sigport.org/3103
Mohammed Bennamoun, Senjian An, Ferdous Sohel, Farid Boussaid, 2018. CLASSIFICATION OF CORALS IN REFLECTANCE AND FLUORESCENCE IMAGES USING CONVOLUTIONAL NEURAL NETWORK REPRESENTATIONS. Available at: http://sigport.org/3103.
Mohammed Bennamoun, Senjian An, Ferdous Sohel, Farid Boussaid. (2018). "CLASSIFICATION OF CORALS IN REFLECTANCE AND FLUORESCENCE IMAGES USING CONVOLUTIONAL NEURAL NETWORK REPRESENTATIONS." Web.
1. Mohammed Bennamoun, Senjian An, Ferdous Sohel, Farid Boussaid. CLASSIFICATION OF CORALS IN REFLECTANCE AND FLUORESCENCE IMAGES USING CONVOLUTIONAL NEURAL NETWORK REPRESENTATIONS [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3103