Vladimir Risojevic

Fusion of Global and Local Descriptors for Remote Sensing Image Classification

Very high resolution remote sensing images offer increased amount of details available for image interpretation. However, despite enhanced resolution, these details result in spectral inhomogeneities, making automated image classification more difficult. In this letter, we propose to combine texture and local image features to address this problem. We first address the enhanced Gabor texture descriptor which is a global descriptor based on cross correlations between subbands and show that it achieves very good results in classification of aerial images showing a single thematic class. Next, the performances obtained on individual land cover/land use classes using our global texture descriptor and local scale-invariant feature transform descriptor are compared. We identify classes of images best suited for each descriptor and argue that these descriptors encode complementary information. Finally, a hierarchical approach for the fusion of global and local descriptors is proposed and evaluated over a number of classifiers. The proposed descriptor fusion approach exhibits significantly improved classification results, reaching the accuracy of around 90%.

Gabor descriptors for aerial image classification

The amount of remote sensed imagery that has become available by far surpasses the possibility of manual analysis. One of the most important tasks in the analysis of remote sensed images is land use classification. This task can be recast as semantic classification of remote sensed images. In this paper we evaluate classifiers for semantic classification of aerial images. The evaluated classifiers are based on Gabor and Gist descriptors which have been long established in image classification tasks. We use support vector machines and propose a kernel well suited for using with Gabor descriptors. These simple classifiers achieve correct classification rate of about 90% on two datasets. From these results follows that, in aerial image classification, simple classifiers give results comparable to more complex approaches, and the pursuit for more advanced solutions should continue having this in mind.

Aerial image classification using structural texture similarity

There is an increasing need for algorithms for automatic analysis of remote sensing images and in this paper we address the problem of semantic classification of aerial images. For the task at hand we propose and evaluate local structural texture descriptor and similarity measure. Nearest neighbor classifier based on the proposed descriptor and similarity measure, as well as image-to-class similarity, improves classification rates over the state-of-the-art on two datasets of aerial images. We evaluate the design choices and show that rich subband statistics, perceptually-based structural texture similarity measure and image-to-class similarity all contribute to the good performance of our classifier.

Unsupervised Quaternion Feature Learning for Remote Sensing Image Classification

Bag-of-words image representations based on local descriptors are common in image classification and retrieval tasks. However, in order to achieve state-of-the-art results, complex hand-crafted feature filters and/or support vector classifiers with nonlinear kernels are needed. Compared with hand-crafted features, unsupervised feature learning is a popular alternative, which results in feature filters adapted to the problem domain at hand. Although both color and intensity are important cues for remote sensing image classification and color images are commonly used for unsupervised feature learning, most of the existing algorithms do not take into account interrelationships between intensity and color information. We address this problem using quaternion representation for color images and propose unsupervised learning of quaternion feature filters, as well as feature encoding using quaternion orthogonal matching pursuit (Q-OMP). By using quaternion representation, we are able to jointly encode intensity and color information in an image. We obtain local descriptors by soft thresholding and computing absolute values of scalar and three vector parts of the quaternion-valued sparse code. Local descriptors are pooled, power-law transformed, and normalized, yielding the resulting image representation. The experimental results on UC Merced Land Use and Brazilian Coffee Scenes datasets are comparable or better than the state of the art, demonstrating the effectiveness of the proposed approach. The proposed method for quaternion feature learning is able to adapt to the characteristics of the available data, and being fully unsupervised, it emerges as a viable alternative to both hand-crafted representations and convolutional neural networks, especially in application scenarios with scarce-labeled training data.

Unsupervised learning of quaternion features for image classification

Unsupervised feature learning is a very popular trend in image classification. Most of the methods for unsupervised feature learning produces filters which operate either on intensity or color information. In this paper we propose a quaternion-based approach for unsupervised feature learning which makes possible joint encoding of the intensity and color information. The image representation is computed using quaternion principal component analysis and k-means clustering. We experimentally show that our approach outperforms the existing approach for unsupervised feature learning from color images, achieving classification accuracy of 91% on a dataset of remote sensing images.

A GPU implementation of a structural-similarity-based aerial-image classification

There is an increasing need for fast and efficient algorithms for the automatic analysis of remote-sensing images. In this paper we address the implementation of the semantic classification of aerial images with general-purpose graphics-processing units (GPGPUs). We propose the calculation of a local Gabor-based structural texture descriptor and a structural texture similarity metric combined with a nearest-neighbor classifier and image-to-class similarity on CUDA supported graphics-processing units. We first present the algorithm and then describe the GPU implementation and optimization with the CUDA programming model. We then evaluate the results of the algorithm on a dataset of aerial images and present the execution times for the sequential and parallel implementations of the whole algorithm as well as measurements only for the selected steps of the algorithm. We show that the algorithms for the image classification can be effectively implemented on the GPUs. In our case, the presented algorithm is around 39 times faster on the Tesla C1060 unit than on the Core i5 650 CPU, while keeping the same success rate of classification.

A simple pipelined squaring circuit for DSP

There are many digital signal processing applications where a shorter time delay of algorithms and efficient implementations are more important than accuracy. Since squaring is one of the fundamental operations widely used in digital signal processing algorithms, approximate squaring is proposed. We present a simple way of approximate squaring that allows achieving a desired accuracy. The proposed method uses the same simple combinational logic for the first approximation and correction terms. Performed analysis for various bit-length operands and level of approximation showed that maximum relative errors and average relative errors decrease significantly by adding more correction terms. The proposed squaring method can be implemented with a great level of parallelism. The pipelined implementation is also proposed in this paper. The proposed squarer achieved significant savings in area and power when compared to multiplier based squarer. As an example, an analysis of the impact of Euclidean distance calculation by approximate squaring on image retrieval is performed.

Evaluation of Bag-of-Colors descriptor for land use classification

The objective of this paper is to evaluate Bag-of-Colors (BoC) descriptor for land use classification. BoC can be used either as a global or local descriptor. In this paper we present and evaluate both approaches. We analyze the influence of different parameters on classification accuracy and introduce a modification of descriptor extraction process, which significantly influences the classification results and performance.

Evaluation of convnets for large-scale scene classification from high-resolution remote sensing images

Convolutional neural networks (convnets) have made possible a number of breakthroughs in image classification and other computer vision problems. However, in order to successfully apply convnets to a new task it should be trained on a large set of labeled samples. Acquisition of a large number of manually labeled remote sensing images requires highly trained analysts which makes it a very expensive task. This is the main reason why we still lack large training sets of remote sensing images. Nevertheless, convnets can still be applied to remote sensing image classification by virtue of using convnets pretrained on another large dataset and fine-tuned to the task at hand. In this paper we investigate the use of pretrained and fine-tuned convnets for both end-to-end classification and feature extraction from remote sensing images. We analyze the quality of the features extracted from various layers of the network from the standpoint of classification accuracy. Using a fine-tuned ResNet we obtain classification accuracy of over 94% on challenging AID dataset.

Splash detection in surveillance videos of offshore fish production plants

Automatic detection of fish welfare related parameters is a very important step in the process of aquaculture production control. Poor handling, and lack of control of the state of the biomass in production plants, may lead to various disease outbreaks, chronic stress and physical trauma, which can influence mortality, which is directly related to profit loss. Automated and objective splash detection provides reliable information about surface activity, which may provide valuable insight into the state of the fish in the cage. In this paper, we propose an algorithm based on Support Vector Machines (SVM), for automatic splash detection in plant surveillance videos, obtained using an unmanned aerial vehicle. We also evaluate the use of Bag-of-Words (BoW) and Vector of Locally Aggregated Descriptors (VLAD) descriptors, for use in splash detection algorithms.