Infrared small target detection using directional highpass filters based on LS-SVM

Target Detection Algorithm Based on Two Layers Human Visual System

“…Various algorithms have been developed in the past few decades [2][3][4]. Conventional small target detection methods such as top-hat filter [2], max-mean/max-median filter [3] and high-pass filters based on least spuares support vector machine (LS-SVM) [4] are widely used to reduce the background clutters.…”

“…Various algorithms have been developed in the past few decades [2][3][4]. Conventional small target detection methods such as top-hat filter [2], max-mean/max-median filter [3] and high-pass filters based on least spuares support vector machine (LS-SVM) [4] are widely used to reduce the background clutters. In recent years, a series of simple and fast algorithm based on Fourier transform was proposed, such as spectral residual (SR) [5], phase spectrum of Fourier transform (PFT) [6], hypercomplex Fourier transform (HFT) [7].…”

Small target detection based on accumulated center-surround difference measure

“…Existing background suppression methods for single-frame infrared image are mainly based on the filtering methods [10,11]. The LS-SVM [11] method uses filter templates, which can suppress most part of the correlative background but may be easily interfered because of the strong fluctuation of background clutters.…”

A simple and efficient saliency extraction method based on multi-scale horizon-directional filter for infrared dim small target detection

“…These properties make small targets highly difficult to be detected. Many algorithms have been proposed to detect infrared small targets, including mathematical morphology based algorithms [9][10][11][12][13][14], filter-based algorithms [6][7][8], wavelet based algorithm [15], machine learning based algorithms [16][17][18] and saliency based algorithms [19][20][21].…”

“…Machine learning based algorithms [16][17][18] are trained to find specific target categories like tank, truck and ship. The drawback is that these algorithms require extensive training and they are not able to cope with objects that do not belong to the predefined object categories.…”