Context This database is intended for experiments in 3D object reocgnition from shape. It contains images of 50 toys belonging to 5 generic categories: **four-legged animals**, **human figures**, **airplanes**, **trucks**, and **cars**. The objects were imaged by 2 cameras under 6 lighting conditions, 9 elevations , and 18 azimuths. The training set is composed of 5 instances of each category (instances 4, 6, 7, 8 and 9), and the test set of the remaining 5 instances (instances 0, 1, 2, 3, and 5). ![Some images from the NORB][1] **The Dataset was created by: Fu Jie Huang, Yann LeCun Courant Institute, New York University October, 2005** The dataset as well as some of this overview was taken from : [The official site][2] Content The files are in a simple binary matrix format, with file postfix ".mat". - The ***"-dat"*** files store the image sequences. Each *"-dat"* file stores 24.300 image **PAIRS** (5 categories, 5 instances, 6 lightings, 9 elevations, and 18 azimuths). **PAIRS :** Each pair is composed of 2 images(24.300 * 2 = 46.600 Total), one left and one right and is commontly used for experiments in binocular mode. For experiments in monocular mode use just one of the two images (24.300 Total). - The ***"-cat"*** files store the corresponding category of the images. The corresponding *"-cat"* file contains 24,300 category labels (**0** for animal, **1** for human, **2** for plane, **3** for truck, **4** for car). - Each ***"-info"*** file stores 24,300 4-dimensional vectors, which contain additional information about the corresponding images: - The instance in the category (0 to 9) - The elevation (0 to 8, which mean cameras are 30, 35,40,45,50,55,60,65,70 degrees from the horizontal respectively) - The azimuth (0,2,4,...,34, multiply by 10 to get the azimuth in degrees) - The lighting condition (0 to 5) For regular training and testing, "-dat" and "-cat" files are sufficient. "-info" files are provided in case some other forms of classification or preprocessing are needed. File Format - The files are stored in the so-called **"binary matrix"** file format, which is a simple format for vectors and multidimensional matrices of various element types. Binary matrix files begin with a file header which describes the type and size of the matrix, and then comes the binary image of the matrix. - The header is best described by a C structure: struct header { int magic; // 4 bytes int ndim; // 4 bytes, little endian int dim[3]; }; *(Note that when the matrix has less than 3 dimensions, say, it's a 1D vector, then dim[1] and dim[2] are both 1. When the matrix has more than 3 dimensions, the header will be followed by further dimension size information. Otherwise, after the file header comes the matrix data, which is stored with the index in the last dimension changes the fastest.)* - Since the files are generated on an Intel machine, they use the **little-endian** scheme to encode the **4-byte integers**. Pay attention when you read the files on machines that use big-endian. - The "-dat" files store a 4D tensor of dimensions 24300x2x96x96. Each files has 24,300 image pairs, (obviously, each pair has 2 images), and each image is 96x96 pixels. - The "-cat" files store a 2D vector of dimension 24,300x1. The "-info" files store a 2D matrix of dimensions 24300x4. **You can find a piece of Matlab code to show how to read an example file at the end of the official website [here][3]** Acknowledgements The Dataset was created by: Fu Jie Huang, Yann LeCun Y. LeCun, F.J. Huang, L. Bottou, Learning Methods for Generic Object Recognition with Invariance to Pose and Lighting. IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR) 2004 Courant Institute, New York University October, 2005 The dataset as well as some of this overview was taken from : [The official site][4] **TERMS / COPYRIGHT** This database is provided for research purposes. It cannot be sold. Publications that include results obtained with this database should reference the following paper: Y. LeCun, F.J. Huang, L. Bottou, Learning Methods for Generic Object Recognition with Invariance to Pose and Lighting. IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR) 2004 Inspiration - What types of machine learning models perform best on this dataset? - Developing learning systems that can recognize generic object purely from their shape, independently of pose, illumination. - Improve the 6% error Rate [According to this results ][5] [1]: https://www.researchgate.net/profile/Sven_Behnke/publication/221080312/figure/fig2/AS:393937547218944@1470933438013/Fig-3-Images-from-the-NORB-normalized-uniform-dataset.ppm [2]: https://cs.nyu.edu/~ylclab/data/norb-v1.0-small/ [3]: https://cs.nyu.edu/~ylclab/data/norb-v1.0-small/ [4]: https://cs.nyu.edu/~ylclab/data/norb-v1.0-small/ [5]: https://cs.nyu.edu/~yann/research/norb/