1. Codes

  • DOTA: A Large-scale dataset for object detection in aerial images [PDF], [codes], [gitHub].
    G.-S. Xia, X. Bai, J. Ding, Z. Zhu, S. Belongie, J. Luo, M. Datcu, M. Pelillo, L. Zhang.
    IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
  • Texture characterization using shape co-occurrence patterns [PDF], [codes].
    G.-S. Xia, G. Liu, X. Bai, L. Zhang.
    IEEE Trans. on Image Processing, Vol.26, No. 10, pp.5005 - 5018, 2017.
  • A generic framework for the structured abstraction of images [PDF] [Code].
    N. Faraj, G.-S. Xia, J. Delon and Y. Gousseau.
    Expressive 2017, July 29-30, Los Angeles, USA, 2017.
  • AID: A benchmark dataset for performance evaluation of aerial scene classification [PDF] [project-page][codes].
    G.-S. Xia, J. Hu, F. Hu, B. Shi, X. Bai, Y. Zhong, L. Zhang.
    IEEE Trans. on Geoscience and Remote Sensing, Vol. 55, No.7, pp.3965 - 3981, 2017.
  • Anisotropic-scale junction detection and matching for indoor images [PDF][Code].
    N. Xue, G.-S. Xia, X. Bai, L. Zhang, W. Shen.
    IEEE Trans. on Image Processing, doi: 10.1109/TIP.2017.2754945, 2017.
  • Texture Synthesis Through Convolutional Neural Networks and Spectrum Constraints [PDF],[Codes](Oral)
    G. Liu and Y. Gousseau, G.-S. Xia.
    IEEE International Conference on Pattern Recognition (ICPR): Cancun, Mexico, 2016.
  • Globally consistent correspondence of multiple features sets using proximal Gauss-Seidel relaxation [PDF] [Code].
    J.-G. Yu, G.-S. Xia, A. Samal, J. Tian.
    Pattern Recognition,Vol. 51, pp. 255–267, Jan. 2016.
  • Accurate junction detection and characterization in natural images [PDF] [Code].
    G.-S. Xia, J. Delon, Y. Gousseau.
    International Journal of Computer Vision (IJCV), Vol.106, No.1, pp.31-56, 2014.
  • Meaningful object segmentation from SAR images via a multi-scale non-local active contour model [PDF],[Codes].
    G.-S. Xia, G. Liu, W. Yang, L. Zhang.
    IEEE Trans. on Geoscience and Remote Sensing, Vol.54, No.4, pp.2108-2123, 2016.
  • Synthesizing and mixing stationary Gaussian texture models [PDF].[Codes]
    G.-S. Xia, S. Ferradans, G. Peyre, J-F. Aujol.
    SIAM Journal on Imaging Science (SIIMS), Vol.7, No.1, pp.476-508, 2014.
  • Shape-based invariant texture indexing [PDF], [Codes].
    G.-S. Xia, J. Delon, Y. Gousseau.
    International Journal of Computer Vision (IJCV), Vol.88, No.3, pp.382-403, 2010.

2. Datasets [webpage]

  • GID Dataset [Download]
  • Gaofen Image Dataset (GID) a large-scale dataset for land use and land cover (LULC) classification. It contains 150 high-quality Gaofen-2 (GF-2) images acquired from more than 60 different cities in China. And these images cover the geographic areas that exceed 50,000 km2. Images in GID have high intra-class diversity coupled with low inter-class separability. Therefore, GID can provide the research community with a high-quality data resource to advance the state-of-the-art in LULC classification.

  • DOTA-v1.0 Dataset [Download], [gitHub]
  • Dota is a large-scale dataset for object detection in aerial images. It can be used to develop and evaluate object detectors in aerial images. We will continue to update DOTA, to grow in size and scope and to reflect evolving real-world conditions. For the DOTA-v1.0, as described in the paper, it contains 2806 aerial images from different sensors and platforms. Each image is of the size in the range from about 800 × 800 to 4000 × 4000 pixels and contains objects exhibiting a wide variety of scales, orientations, and shapes. These DOTA images are then annotated by experts in aerial image interpretation using 15 common object categories. The fully annotated DOTA images contains 188, 282 instances, each of which is labeled by an arbitrary (8 d.o.f.) quadrilateral.
    For more details, refer to the arXiv preprint of DOTA.

  • AID Dataset [Download] [gitHub]
  • The images in AID are actually multi-source, as Google Earth images are from different remote imaging sensors. AID dataset has a number of 10000 images within 30 classes. AID has multi-resolutions: the pixel-resolution changes from about 8 meters to about half a meter, and thus the size of each aerial image is fixed to be 600 × 600 pixels to cover a scene with various resolution.

    -G.-S. Xia, J. Hu, F. Hu, B. Shi, X. Bai, Y. Zhong, L. Zhang, X. Lu, “AID: A benchmark dataset for performance evaluation of aerial scene classification”, IEEE Transactions on Geoscience and Remote Sensing, vol. 55, no. 7, pp. 3965-3981, 2017.[PDF].
  • WHU-RS19 Datasets [Download]
  • WHU-RS19 is a dataset of remote sensing images collected from Google Earth, it can be used for scene classification and retrieval.

    -G.-S. Xia, W. Yang, J. Delon, Y. Gousseau. H. Maitre, H. Sun, "Structural high-resolution satellite image indexing". Symposium: 100 Years ISPRS - Advancing Remote Sensing Science: Vienna, Austria, 2010