Do deep features generalize from everyday objects to remote sensing and aerial scenes domains?

Penatti, Otávio Augusto Bizetto; Nogueira, Keiller; Santos, Jefersson A. dos

doi:10.1109/cvprw.2015.7301382

Cited by 591 publications

(408 citation statements)

References 39 publications

(61 reference statements)

Supporting

Mentioning

402

Contrasting

Unclassified

Order By: Relevance

“…AlexNet and VGG-16 are bigger CNN and winner of the ImageNet competition in 2012 and 2014. Preliminary experiments show that the CNN overall accuracy is improved by 10% when using pre-trained weights on the ImageNet dataset, which is consistent with results from [6,11]. Therefore, those CNN will be simply fine-tuned with input patches at resolution 224 × 224 and 227 × 227, respectively.…”

Section: Learning a Vehicle Classifiersupporting

confidence: 49%

“…For example, FCN-based models are now the state-of-the-art on the ISPRS Vaihingen Semantic Labeling dataset [9,10]. Therefore, even though remote sensing images do not share the same structure as natural images, traditional computer vision deep networks are able to successfully extract semantics from them, which was already known for deep CNN-based classifiers [11]. This encourages us to investigate further: can we use deep networks to tackle an especially hard remote sensing task, namely object segmentation?…”

Section: Introductionmentioning

confidence: 99%

“…For example, Penatti, O.A.B. et al [11] examined transfer learning from pre-trained convolutional neural networks (CNN) on traditional red-green-blue (RGB) images to remote sensing data. This work has been later consolidated by [6] to better understand CNN-based classification of Earth Observation images.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Segment-before-Detect: Vehicle Detection and Classification through Semantic Segmentation of Aerial Images

2017

View full text Add to dashboard Cite

Like computer vision before, remote sensing has been radically changed by the introduction of deep learning and, more notably, Convolution Neural Networks. Land cover classification, object detection and scene understanding in aerial images rely more and more on deep networks to achieve new state-of-the-art results. Recent architectures such as Fully Convolutional Networks can even produce pixel level annotations for semantic mapping. In this work, we present a deep-learning based segment-before-detect method for segmentation and subsequent detection and classification of several varieties of wheeled vehicles in high resolution remote sensing images. This allows us to investigate object detection and classification on a complex dataset made up of visually similar classes, and to demonstrate the relevance of such a subclass modeling approach. Especially, we want to show that deep learning is also suitable for object-oriented analysis of Earth Observation data as effective object detection can be obtained as a byproduct of accurate semantic segmentation. First, we train a deep fully convolutional network on the ISPRS Potsdam and the NZAM/ONERA Christchurch datasets and show how the learnt semantic maps can be used to extract precise segmentation of vehicles. Then, we show that those maps are accurate enough to perform vehicle detection by simple connected component extraction. This allows us to study the repartition of vehicles in the city. Finally, we train a Convolutional Neural Network to perform vehicle classification on the VEDAI dataset, and transfer its knowledge to classify the individual vehicle instances that we detected.

show abstract

Section: Learning a Vehicle Classifiersupporting

confidence: 49%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Segment-before-Detect: Vehicle Detection and Classification through Semantic Segmentation of Aerial Images

2017

View full text Add to dashboard Cite

show abstract

“…Previous studies reported the superiority of fine-tuning relative to full-training for classification of very high resolution aerial imagery, although full-training was found to be more accurate relative to fine-tuning for classification of multi-spectral satellite data [26,45]. In particular, Nogueira et al (2017) evaluated the efficiency of fine-tuning and full-training strategies of some well-known deep CNNs (e.g., AlexNet and GoogLeNet) for classification of three well-known datasets, including UCMerced land-use [46], RS19 dataset [47], and Brazilian Coffee Scenes [48]. The fine-tuning strategy yielded a higher accuracy for the first two datasets, likely due to their similarity with the ImageNet dataset, which was originally used for training deep CNNs.…”

Section: Figure 10mentioning

confidence: 99%

Very Deep Convolutional Neural Networks for Complex Land Cover Mapping Using Multispectral Remote Sensing Imagery

et al. 2018

View full text Add to dashboard Cite

Despite recent advances of deep Convolutional Neural Networks (CNNs) in various computer vision tasks, their potential for classification of multispectral remote sensing images has not been thoroughly explored. In particular, the applications of deep CNNs using optical remote sensing data have focused on the classification of very high-resolution aerial and satellite data, owing to the similarity of these data to the large datasets in computer vision. Accordingly, this study presents a detailed investigation of state-of-the-art deep learning tools for classification of complex wetland classes using multispectral RapidEye optical imagery. Specifically, we examine the capacity of seven well-known deep convnets, namely DenseNet121, InceptionV3, VGG16, VGG19, Xception, ResNet50, and InceptionResNetV2, for wetland mapping in Canada. In addition, the classification results obtained from deep CNNs are compared with those based on conventional machine learning tools, including Random Forest and Support Vector Machine, to further evaluate the efficiency of the former to classify wetlands. The results illustrate that the full-training of convnets using five spectral bands outperforms the other strategies for all convnets. InceptionResNetV2, ResNet50, and Xception are distinguished as the top three convnets, providing state-of-the-art classification accuracies of 96.17%, 94.81%, and 93.57%, respectively. The classification accuracies obtained using Support Vector Machine (SVM) and Random Forest (RF) are 74.89% and 76.08%, respectively, considerably inferior relative to CNNs. Importantly, InceptionResNetV2 is consistently found to be superior compared to all other convnets, suggesting the integration of Inception and ResNet modules is an efficient architecture for classifying complex remote sensing scenes such as wetlands.

show abstract

“…In fact, spatial adjacent pixels usually share similar spectral characteristics and have the same label, and using spatial information can reduce the uncertainty of samples and suppress the salt-and-pepper noise of classification results [57]. Furthermore, recent studies have found that a higher layer of the deep hierarchical model produces increasingly abstract representations and is increasingly invariant to some transformations [58,59]. Consequently, we present a new framework that jointly learns the spectral-spatial features via a deep hierarchical architecture.…”

Section: Spectral-spatial Responsementioning

confidence: 99%

Spectral-Spatial Response for Hyperspectral Image Classiﬁcation

Wei

Zhou

2017

Remote Sensing

View full text Add to dashboard Cite

This paper presents a hierarchical deep framework called Spectral-Spatial Response (SSR) to jointly learn spectral and spatial features of Hyperspectral Images (HSIs) by iteratively abstracting neighboring regions. SSR forms a deep architecture and is able to learn discriminative spectral-spatial features of the input HSI at different scales. It includes several existing spectral-spatial-based methods as special scenarios within a single unified framework. Based on SSR, we further propose the Subspace Learning-based Networks (SLN) as an example of SSR for HSI classification. In SLN, the joint spectral and spatial features are learned using templates simply learned by Marginal Fisher Analysis (MFA) and Principal Component Analysis (PCA). An important contribution to the success of SLN is the exploitation of label information of training samples and the local spatial structure of HSI. Extensive experimental results on four challenging HSI datasets taken from the Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) and Reflective Optics System Imaging Spectrometer (ROSIS) airborne sensors show the implementational simplicity of SLN and verify the superiority of SSR for HSI classification.

show abstract

Do deep features generalize from everyday objects to remote sensing and aerial scenes domains?

Cited by 591 publications

References 39 publications

Segment-before-Detect: Vehicle Detection and Classification through Semantic Segmentation of Aerial Images

Segment-before-Detect: Vehicle Detection and Classification through Semantic Segmentation of Aerial Images

Very Deep Convolutional Neural Networks for Complex Land Cover Mapping Using Multispectral Remote Sensing Imagery

Spectral-Spatial Response for Hyperspectral Image Classiﬁcation

Contact Info

Product

Resources

About