Multi-channel descriptors and ensemble of Extreme Learning Machines for classification of remote sensing images

Cvetković, Stevica; Stojanović, Miloš B.; Nikolić, Saša V.

doi:10.1016/j.image.2015.09.004

Cited by 11 publications

(7 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The reasons are that images of beach and forest present rich texture and structures information; within-class similarity for the beach and forest categories is high but relatively low for category of storage tank; and some images of storage tank are similar to other class such as buildings. Method Accuracy(Mean ± std) BOVW [28] 76.8 SPM [28] 75.3 BOVW + Spatial Co-occurrence Kernel [28] 77.7 Color Gabor [28] 80.5 Color histogram (HLS) [28] 81.2 Structural texture similarity [7] 86.0 Unsupervised feature learning [33] 81.7 ± 1.2 Saliency-Guided unsupervised feature learning [34] 82.7 ± 1.2 Concentric circle-structured multiscale BOVW [5] 86.6 ± 0.8 Multifeature concatenation [35] 89.5 ± 0.8 Pyramid-of-Spatial-Relatons (PSR) [36] 89.1 MCBGP + E-ELM [37] 86.52 ± 1.3 ConvNet with specific spatial features [38] 89.39 ± 1.10 gradient boosting randomconvolutional network [39] 94.53 GoogLeNet [40] 92.80 ± 0.61 OverFeatConvNets [40] 90.91 ± 1.19 MS-CLBP [17] 90.6 ± 1. When compared with CNNs, it can be found that the classification accuracy of CNNs is close to that of our method.…”

Section: Methodsmentioning

confidence: 99%

Remote Sensing Image Scene Classification Using Multi-Scale Completed Local Binary Patterns and Fisher Vectors

et al. 2016

View full text Add to dashboard Cite

An effective remote sensing image scene classification approach using patch-based multi-scale completed local binary pattern (MS-CLBP) features and a Fisher vector (FV) is proposed. The approach extracts a set of local patch descriptors by partitioning an image and its multi-scale versions into dense patches and using the CLBP descriptor to characterize local rotation invariant texture information. Then, Fisher vector encoding is used to encode the local patch descriptors (i.e., patch-based CLBP features) into a discriminative representation. To improve the discriminative power of feature representation, multiple sets of parameters are used for CLBP to generate multiple FVs that are concatenated as the final representation for an image. A kernel-based extreme learning machine (KELM) is then employed for classification. The proposed method is extensively evaluated on two public benchmark remote sensing image datasets (i.e., the 21-class land-use dataset and the 19-class satellite scene dataset) and leads to superior classification performance (93.00% for the 21-class dataset with an improvement of approximately 3% when compared with the state-of-the-art MS-CLBP and 94.32% for the 19-class dataset with an improvement of approximately 1%).

show abstract

Section: Methodsmentioning

confidence: 99%

Remote Sensing Image Scene Classification Using Multi-Scale Completed Local Binary Patterns and Fisher Vectors

et al. 2016

View full text Add to dashboard Cite

show abstract

“…Ensembles of ELMs are introduced to overcome known drawbacks of ELM caused by the randomness of input weights and biases [1,9]. In order to improve accuracy of an ensemble, we proposed to directly accumulate output scores of the ELMs into a final decision [4], instead of accumulating their binary votes. The proposed algorithm, denoted as Average Score Aggregation (ASA), is described below.…”

Section: Ensemble Of Elms (E-elm) Using Average Score Aggregationmentioning

confidence: 99%

“…Classification accuracy is measured by varying number of hidden neurons per ELM for a fixed number of ELMs in ensemble (k = 15). We used ASA strategy [4] as aggregation method. Results are presented in Figure 3.…”

Section: Algorithm 2 E-elm Testing Phasementioning

confidence: 99%

“…For aggregating multiple ELMs into a single classification system, the most common method is voting based aggregation [6,9]. Recently we introduced Average Score Aggregation [4] which shows improved classification accuracy without additional complexity. In this paper we will apply Average Score Aggregation to the ELM ensembles for image classification.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Real-time image classification using LBP and ensembles of ELM

Cvetković¹,

Rajković²,

Nikolić³

2016

Sci Publ State Univ Novi Pazar Ser A

View full text Add to dashboard Cite

This paper presents a method for real-time image classification which combines Local Binary Pattern (LBP) descriptor and ensembles of Extreme Learning Machines (ELM). We start by extraction of a multi-channel LBP descriptor over separate color channels and several scales of the input image. The descriptor is characterized by compactness and robustness to illumination and resolution changes. Image category prediction is done using recently introduced specific single layer neural network called Extreme Learning Machine (ELM). To overcome main disadvantages of the ELM, instability and non-optimal output, we combined multiple ELMs into an ensemble using appropriate score aggregation strategy. Our evaluation on a standard benchmark dataset consisting of a thousand images from ten categories, has shown high accuracy of results while executing in real-time during tests.

show abstract

“…By extracting features of the input data from the bottom to the top of the network, deep-learning models can form the high-level abstract features suitable for pattern classification [26]. Among numerous deep-learning models, a convolutional neural network (CNN) has a relatively small number of weights owing to local connections and sharing weights [27]. Moreover, the multidimensional tensor data, for instance HSIs, can be directly input into 3D convolutional neural networks (3D-CNNs), which helps to preserve the original relevant information of the data and avoids complex data reconstruction [28][29][30].…”

Section: Introductionmentioning

confidence: 99%

Hyperspectral Image Classification Based on Parameter-Optimized 3D-CNNs Combined with Transfer Learning and Virtual Samples

Sun

Yue

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

Recent research has shown that spatial-spectral information can help to improve the classification of hyperspectral images (HSIs). Therefore, three-dimensional convolutional neural networks (3D-CNNs) have been applied to HSI classification. However, a lack of HSI training samples restricts the performance of 3D-CNNs. To solve this problem and improve the classification, an improved method based on 3D-CNNs combined with parameter optimization, transfer learning, and virtual samples is proposed in this paper. Firstly, to optimize the network performance, the parameters of the 3D-CNN of the HSI to be classified (target data) are adjusted according to the single variable principle. Secondly, in order to relieve the problem caused by insufficient samples, the weights in the bottom layers of the parameter-optimized 3D-CNN of the target data can be transferred from another well trained 3D-CNN by a HSI (source data) with enough samples and the same feature space as the target data. Then, some virtual samples can be generated from the original samples of the target data to further alleviate the lack of HSI training samples. Finally, the parameter-optimized 3D-CNN with transfer learning can be trained by the training samples consisting of the virtual and the original samples. Experimental results on real-world hyperspectral satellite images have shown that the proposed method has great potential prospects in HSI classification.

show abstract

Multi-channel descriptors and ensemble of Extreme Learning Machines for classification of remote sensing images

Cited by 11 publications

References 32 publications

Remote Sensing Image Scene Classification Using Multi-Scale Completed Local Binary Patterns and Fisher Vectors

Remote Sensing Image Scene Classification Using Multi-Scale Completed Local Binary Patterns and Fisher Vectors

Real-time image classification using LBP and ensembles of ELM

Hyperspectral Image Classification Based on Parameter-Optimized 3D-CNNs Combined with Transfer Learning and Virtual Samples

Contact Info

Product

Resources

About