A huge effort has been applied in image classification to create high quality thematic maps and to establish precise inventories about land cover use. The peculiarities of Remote Sensing Images (RSIs) combined with the traditional image classification challenges made RSIs classification a hard task. Our aim is to propose a kind of boost-classifier adapted to multi-scale segmentation. We use the paradigm of boosting, whose principle is to combine weak classifiers to build an efficient global one. Each weak classifier is trained for one level of the segmentation and one region descriptor. We have proposed and tested weak classifiers based on linear SVM and region distances provided by descriptors. The experiments were performed on a large image of coffee plantations. We have shown in this paper that our approach based on boosting can detect the scale and set of features best suited to a particular training set. We have also shown that hierarchical multi-scale analysis is able to reduce training time and to produce a stronger classifier. We compare the proposed methods with a baseline based on SVM with RBF kernel. The results show that the proposed methods outperform the baseline.
Abstract. This paper presents a novel re-ranking approach based on contextual information used to improve the effectiveness of ContentBased Image Retrieval (CBIR) tasks. In our approach, image processing techniques are applied to ranked lists defined by CBIR descriptors. Conducted experiments involving shape, color, and texture descriptors demonstrate the effectiveness of our method.
Generalized assessments of the accuracy of spectroscopic estimates of ecologically important leaf traits, such as leaf mass per area (LMA) and leaf dry matter content (LDMC), are still lacking for most ecosystems and particularly for non-forested and/or seasonally dry tropical vegetation. Here, we tested the ability of using leaf reflectance spectra to estimate LMA and LDMC and classify plant growth forms within the cerrado and campo rupestre vegetation, a seasonally dry non-forest vegetation types of Southeastern Brazil, filling an existing gap in published assessments of leaf optical properties and plant traits in such environments. We measured leaf reflectance spectra from 1648 individual plants comprising grasses, herbs, shrubs, and trees, developed partial least squares regression (PLSR) models linking LMA and LDMC to leaf spectra (400-2500 nm), and identified the spectral regions with the greatest discriminatory power among growth forms using Bhattacharyya distances. We accurately predicted leaf functional traits and identified different growth forms. LMA was overall more accurately predicted (RMSE = 8.58%) than LDMC (RMSE = 9.75%). Our model including all sampled plants was not biased towards any particular growth form, but growth-form specific models yielded higher accuracies and showed that leaf traits from woody plants can be more accurately estimated than for grasses and forbs, independently of the trait measured. We observed a large range of LMA values (31.80 -620.81 g/m 2 ), rarely observed in tropical or temperate forests, and demonstrated that values above 300 g/m 2 cannot be accurately estimated. Our results suggest that spectroscopy may have an intrinsic saturation point, and/or that PLSR, the current approach of choice for estimating traits from plant spectra, is not able to model the entire range of LMA values. This finding has very important implications to our ability to use field, airborne, and orbital spectroscopic methods to derive generalizable functional information. We thus highlight the need for increasing spectroscopic sampling and research efforts in drier non-forested environments, where environmental pressures lead to leaf adaptations and allocation strategies that are very different from forested ecosystems, producing thicker leaves. Our findings also confirm that leaf reflectance spectra can provide important information regarding differences in leaf metabolism, structure, and chemical composition. Such information enabled us to accurately discriminate plant growth forms in these environments regardless the lack of variation in leaf economics traits, encouraging further adoption of remote sensing methods by ecologists and allowing a more comprehensive assessment of plant functional diversity.
In this paper, we explore transformed spaces, represented by image illuminant maps, to propose a methodology for selecting complementary forms of characterizing visual properties for an effective and automated detection of image forgeries. We combine statistical telltales provided by different image descriptors that explore color, shape, and texture features. We focus on detecting image forgeries containing people and present a method for locating the forgery, specifically, the face of a person in an image. Experiments performed on three different open-access data sets show the potential of the proposed method for pinpointing image forgeries containing people. In the two first data sets (DSO-1 and DSI-1), the proposed method achieved a classification accuracy of 94% and 84%, respectively, a remarkable improvement when compared with the state-of-the-art methods. Finally, when evaluating the third data set comprising questioned images downloaded from the Internet, we also present a detailed analysis of target images.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.