Published on Mon Jun 20 2016

Multiple Instance Hyperspectral Target Characterization

Alina Zare, Changzhe Jiao, Taylor Glenn

MI-SMF and MI-ACE estimate a discriminative target signature from imprecisely-labeled and mixed training data. Results show progressivelyimproved, consistent performance over existing multiple instance concept learning methods.

0
0
0
Abstract

In this paper, two methods for multiple instance target characterization, MI-SMF and MI-ACE, are presented. MI-SMF and MI-ACE estimate a discriminative target signature from imprecisely-labeled and mixed training data. In many applications, such as sub-pixel target detection in remotely-sensed hyperspectral imagery, accurate pixel-level labels on training data is often unavailable and infeasible to obtain. Furthermore, since sub-pixel targets are smaller in size than the resolution of a single pixel, training data is comprised only of mixed data points (in which target training points are mixtures of responses from both target and non-target classes). Results show improved, consistent performance over existing multiple instance concept learning methods on several hyperspectral sub-pixel target detection problems.

Sat Sep 07 2019
Machine Learning
Multi-Target Multiple Instance Learning for Hyperspectral Target Detection
The proposed methods address the problems above by considering the multiple-instance, imprecisely labeled dataset. They learn a dictionary of target signatures that optimizes detection against a background using the Adaptive Cosine Estimator (ACE) and Spectral Match Filter (SMF)
0
0
0
Tue Oct 31 2017
Computer Vision
Multiple Instance Hybrid Estimator for Hyperspectral Target Characterization and Sub-pixel Target Detection
The Multiple Instance Hybrid Estimator for discriminative target characterization from imprecisely labeled hyperspectral data is presented. The proposed approach addresses these problems by introducing a data mixing model and optimizing the response of the detector.
0
0
0
Mon Jan 09 2017
Computer Vision
Multiple Instance Hybrid Estimator for Learning Target Signatures
Signature-based detectors for hyperspectral target detection rely on knowing the specific target signature in advance. The proposed approach maximizes the response of the hybrid sub-pixel detector within a multiple instancelearning framework and estimates a set of discriminative targets.
0
0
0
Mon Mar 21 2016
Computer Vision
Instance Influence Estimation for Hyperspectral Target Signature Characterization using Extended Functions of Multiple Instances
The Extended Functions of Multiple Instances (eFUMI) algorithm is a generalization of Multiple instance Learning (MIL) In eFUMi, only bag level labels are needed to estimate target signatures from mixed data.
0
0
0
Fri Oct 30 2015
Computer Vision
Estimating Target Signatures with Diverse Density
Hyperspectral target detection algorithms rely on knowing the desired target signature in advance. obtaining an effective target signature can be difficult. The proposed approach addresses uncertainty and imprecision in groundtruth in training data.
0
0
0
Mon Mar 18 2019
Machine Learning
Learning with Sets in Multiple Instance Regression Applied to Remote Sensing
In this paper, we propose a novel approach to tackle the multiple instanceRegression (MIR) problem. This problem arises when the data is a collection of bags, where each bag is made of multiple instances corresponding to the same real-valued label. Our goal is to train a regression model which
0
0
0