Fully Convolutional Network Model Research Articles

An Improved Image Semantic Segmentation Method Based on Superpixels and Conditional Random Fields

This paper proposed an improved image semantic segmentation method based on superpixels and conditional random fields (CRFs). The proposed method can take full advantage of the superpixel edge information and the constraint relationship among different pixels. First, we employ fully convolutional networks (FCN) to obtain pixel-level semantic features and utilize simple linear iterative clustering (SLIC) to generate superpixel-level region information, respectively. Then, the segmentation results of image boundaries are optimized by the fusion of the obtained pixel-level and superpixel-level results. Finally, we make full use of the color and position information of pixels to further improve the semantic segmentation accuracy using the pixel-level prediction capability of CRFs. In summary, this improved method has advantages both in terms of excellent feature extraction capability and good boundary adherence. Experimental results on both the PASCAL VOC 2012 dataset and the Cityscapes dataset show that the proposed method can achieve significant improvement of segmentation accuracy in comparison with the traditional FCN model.

A Semantic Segmentation Algorithm Using FCN with Combination of BSLIC

An image semantic segmentation algorithm using fully convolutional network (FCN) integrated with the recently proposed simple linear iterative clustering (SLIC) that is based on boundary term (BSLIC) is developed. To improve the segmentation accuracy, the developed algorithm combines the FCN semantic segmentation results with the superpixel information acquired by BSLIC. During the combination process, the superpixel semantic annotation is newly introduced and realized by the four criteria. The four criteria are used to annotate a superpixel region, according to FCN semantic segmentation result. The developed algorithm can not only accurately identify the semantic information of the target in the image, but also achieve a high accuracy in the positioning of small edges. The effectiveness of our algorithm is evaluated on the dataset PASCAL VOC 2012. Experimental results show that the developed algorithm improved the target segmentation accuracy in comparison with the traditional FCN model. With the BSLIC superpixel information that is involved, the proposed algorithm can get 3.86%, 1.41%, and 1.28% improvement in pixel accuracy (PA) over FCN-32s, FCN-16s, and FCN-8s, respectively.

Coarse-to-fine salient object detection based on deep convolutional neural networks

With explosive growth of image data, automatic image interpretation becomes more and more important. Saliency detection is one of the fundamental problems. To predict the saliency map, traditional saliency detection approaches use handcrafted features, which are not robust for complex scene. Recently, convolutional neural network (CNN) have shown good performance in computer vision problems. In this paper, we propose a coarse-to-fine approach combining pixel-wise FCN with superpixel-based CNN for detecting salient objects with precise boundaries. Firstly, the fully convolutional network (FCN) model is used to produce a coarse saliency map. Instead of patch-based CNN taking in overlapping patches as samples, the FCN model adopts the pixel-wise structure which can predict the location of the salient objects from the global aspect. Then, superpixel clustering is presented to decompose the image into homogeneous superpixels. For each superpixel, the local superpixel-based CNN model is created to integrate the coarse saliency map with the original image information for refining the detected salient objects with precise boundaries. Experimental results on large benchmark databases demonstrate the proposed method perform well when tested against the state-of-the-art methods.

Automatic Building Segmentation of Aerial Imagery Using Multi-Constraint Fully Convolutional Networks

Automatic building segmentation from aerial imagery is an important and challenging task because of the variety of backgrounds, building textures and imaging conditions. Currently, research using variant types of fully convolutional networks (FCNs) has largely improved the performance of this task. However, pursuing more accurate segmentation results is still critical for further applications such as automatic mapping. In this study, a multi-constraint fully convolutional network (MC–FCN) model is proposed to perform end-to-end building segmentation. Our MC–FCN model consists of a bottom-up/top-down fully convolutional architecture and multi-constraints that are computed between the binary cross entropy of prediction and the corresponding ground truth. Since more constraints are applied to optimize the parameters of the intermediate layers, the multi-scale feature representation of the model is further enhanced, and hence higher performance can be achieved. The experiments on a very-high-resolution aerial image dataset covering 18 km 2 and more than 17,000 buildings indicate that our method performs well in the building segmentation task. The proposed MC–FCN method significantly outperforms the classic FCN method and the adaptive boosting method using features extracted by the histogram of oriented gradients. Compared with the state-of-the-art U–Net model, MC–FCN gains 3.2% (0.833 vs. 0.807) and 2.2% (0.893 vs. 0.874) relative improvements of Jaccard index and kappa coefficient with the cost of only 1.8% increment of the model-training time. In addition, the sensitivity analysis demonstrates that constraints at different positions have inconsistent impact on the performance of the MC–FCN.

3-D Fully Convolutional Networks for Multimodal Isointense Infant Brain Image Segmentation.

Accurate segmentation of infant brain images into different regions of interest is one of the most important fundamental steps in studying early brain development. In the isointense phase (approximately 6-8 months of age), white matter and gray matter exhibit similar levels of intensities in magnetic resonance (MR) images, due to the ongoing myelination and maturation. This results in extremely low tissue contrast and thus makes tissue segmentation very challenging. Existing methods for tissue segmentation in this isointense phase usually employ patch-based sparse labeling on single modality. To address the challenge, we propose a novel 3-D multimodal fully convolutional network (FCN) architecture for segmentation of isointense phase brain MR images. Specifically, we extend the conventional FCN architectures from 2-D to 3-D, and, rather than directly using FCN, we intuitively integrate coarse (naturally high-resolution) and dense (highly semantic) feature maps to better model tiny tissue regions, in addition, we further propose a transformation module to better connect the aggregating layers; we also propose a fusion module to better serve the fusion of feature maps. We compare the performance of our approach with several baseline and state-of-the-art methods on two sets of isointense phase brain images. The comparison results show that our proposed 3-D multimodal FCN model outperforms all previous methods by a large margin in terms of segmentation accuracy. In addition, the proposed framework also achieves faster segmentation results compared to all other methods. Our experiments further demonstrate that: 1) carefully integrating coarse and dense feature maps can considerably improve the segmentation performance; 2) batch normalization can speed up the convergence of the networks, especially when hierarchical feature aggregations occur; and 3) integrating multimodal information can further boost the segmentation performance.

Effective Fusion of Multi-Modal Remote Sensing Data in a Fully Convolutional Network for Semantic Labeling

In recent years, Fully Convolutional Networks (FCN) have led to a great improvement of semantic labeling for various applications including multi-modal remote sensing data. Although different fusion strategies have been reported for multi-modal data, there is no in-depth study of the reasons of performance limits. For example, it is unclear, why an early fusion of multi-modal data in FCN does not lead to a satisfying result. In this paper, we investigate the contribution of individual layers inside FCN and propose an effective fusion strategy for the semantic labeling of color or infrared imagery together with elevation (e.g., Digital Surface Models). The sensitivity and contribution of layers concerning classes and multi-modal data are quantified by recall and descent rate of recall in a multi-resolution model. The contribution of different modalities to the pixel-wise prediction is analyzed explaining the reason of the poor performance caused by the plain concatenation of different modalities. Finally, based on the analysis an optimized scheme for the fusion of layers with image and elevation information into a single FCN model is derived. Experiments are performed on the ISPRS Vaihingen 2D Semantic Labeling dataset (infrared and RGB imagery as well as elevation) and the Potsdam dataset (RGB imagery and elevation). Comprehensive evaluations demonstrate the potential of the proposed approach.

Learning normalized inputs for iterative estimation in medical image segmentation.

In this paper, we introduce a simple, yet powerful pipeline for medical image segmentation that combines Fully Convolutional Networks (FCNs) with Fully Convolutional Residual Networks (FC-ResNets). We propose and examine a design that takes particular advantage of recent advances in the understanding of both Convolutional Neural Networks as well as ResNets. Our approach focuses upon the importance of a trainable pre-processing when using FC-ResNets and we show that a low-capacity FCN model can serve as a pre-processor to normalize medical input data. In our image segmentation pipeline, we use FCNs to obtain normalized images, which are then iteratively refined by means of a FC-ResNet to generate a segmentation prediction. As in other fully convolutional approaches, our pipeline can be used off-the-shelf on different image modalities. We show that using this pipeline, we exhibit state-of-the-art performance on the challenging Electron Microscopy benchmark, when compared to other 2D methods. We improve segmentation results on CT images of liver lesions, when contrasting with standard FCN methods. Moreover, when applying our 2D pipeline on a challenging 3D MRI prostate segmentation challenge we reach results that are competitive even when compared to 3D methods. The obtained results illustrate the strong potential and versatility of the pipeline by achieving accurate segmentations on a variety of image modalities and different anatomical regions.

Classification for High Resolution Remote Sensing Imagery Using a Fully Convolutional Network

As a variant of Convolutional Neural Networks (CNNs) in Deep Learning, the Fully Convolutional Network (FCN) model achieved state-of-the-art performance for natural image semantic segmentation. In this paper, an accurate classification approach for high resolution remote sensing imagery based on the improved FCN model is proposed. Firstly, we improve the density of output class maps by introducing Atrous convolution, and secondly, we design a multi-scale network architecture by adding a skip-layer structure to make it capable for multi-resolution image classification. Finally, we further refine the output class map using Conditional Random Fields (CRFs) post-processing. Our classification model is trained on 70 GF-2 true color images, and tested on the other 4 GF-2 images and 3 IKONOS true color images. We also employ object-oriented classification, patch-based CNN classification, and the FCN-8s approach on the same images for comparison. The experiments show that compared with the existing approaches, our approach has an obvious improvement in accuracy. The average precision, recall, and Kappa coefficient of our approach are 0.81, 0.78, and 0.83, respectively. The experiments also prove that our approach has strong applicability for multi-resolution image classification.