Adaptive Feature Recombination and Recalibration for Semantic Segmentation With Fully Convolutional Networks

Autor: Adriano Pinto, Sérgio Pereira, Victor Alves, Carlos A. Silva, Joana Amorim, Alexandrine Ribeiro
Přispěvatelé: Universidade do Minho
Rok vydání: 2019
Předmět:
FOS: Computer and information sciences
Computer science
Computer Vision and Pattern Recognition (cs.CV)
Computer Science - Computer Vision and Pattern Recognition
recalibration
computer.software_genre
Convolutional neural network
030218 nuclear medicine & medical imaging
03 medical and health sciences
0302 clinical medicine
Segmentation
Deep Learning
Discriminative model
Voxel
FOS: Electrical engineering
electronic engineering
information engineering

Image Processing
Computer-Assisted

Humans
Electrical and Electronic Engineering
fully convolutional network
Science & Technology
Radiological and Ultrasound Technology
Contextual image classification
business.industry
Brain Neoplasms
Image and Video Processing (eess.IV)
adaptive
deep learning
Brain
Pattern recognition
Engenharia Eletrotécnica
Eletrónica e Informática [Engenharia e Tecnologia]

Electrical Engineering and Systems Science - Image and Video Processing
Magnetic Resonance Imaging
recombination
Computer Science Applications
Semantics
Stroke
Kernel (image processing)
Artificial intelligence
Neural Networks
Computer

business
computer
Software
Engenharia e Tecnologia::Engenharia Eletrotécnica
Eletrónica e Informática
Zdroj: Repositório Científico de Acesso Aberto de Portugal
Repositório Científico de Acesso Aberto de Portugal (RCAAP)
instacron:RCAAP
ISSN: 1558-254X
Popis: Fully convolutional networks have been achieving remarkable results in image semantic segmentation, while being efficient. Such efficiency results from the capability of segmenting several voxels in a single forward pass. So, there is a direct spatial correspondence between a unit in a feature map and the voxel in the same location. In a convolutional layer, the kernel spans over all channels and extracts information from them. We observe that linear recombination of feature maps by increasing the number of channels followed by compression may enhance their discriminative power. Moreover, not all feature maps have the same relevance for the classes being predicted. In order to learn the inter-channel relationships and recalibrate the channels to suppress the less relevant ones, squeeze and excitation blocks were proposed in the context of image classification with convolutional neural networks. However, this is not well adapted for segmentation with fully convolutional networks since they segment several objects simultaneously, hence a feature map may contain relevant information only in some locations. In this paper, we propose recombination of features and a spatially adaptive recalibration block that is adapted for semantic segmentation with fully convolutional networks- the SegSE block. Feature maps are recalibrated by considering the cross-channel information together with spatial relevance. The experimental results indicate that recombination and recalibration improve the results of a competitive baseline, and generalize across three different problems: brain tumor segmentation, stroke penumbra estimation, and ischemic stroke lesion outcome prediction. The obtained results are competitive or outperform the state of the art in the three applications.
This work was supported in part by the Fundacao para a Ciencia e Tecnologia (FCT) with the reference project UID/EEA/04436/2013, and in part by the COMPETE 2020 with the code POCI-01-0145-FEDER-006941 and COMPETE: POCI-01-0145-FEDER-007043, and in part by the FCT within the Project Scope: UID/CEC/00319/2013. The work of S. Pereira was supported by the FCT, Portugal, under Scholarship PD/BD/105803/2014.
Databáze: OpenAIRE