Registration   Download   Introduction   Data   Task   Development kit   Timetable   Features   Submission   Organizers   Contact



The goal of this competition is to estimate the content of photographs for the purpose of retrieval and automatic annotation using a subset of the large hand-labeled ImageNet dataset (10,000,000 labeled images depicting 10,000+ object categories) as training. Test images will be presented with no initial annotation -- no segmentation or labels -- and algorithms will have to produce labelings specifying what objects are present in the images. In this initial version of the competition, the goal is only to identify the main objects present in images, not to specify the location of objects.


The validation and test data for this competition will consist of 200,000 photographs, collected from flickr and other search engines, hand labeled with the presence or absence of 1000 object categories. The 1000 object categories contain both internal nodes and leaf nodes of ImageNet, but do not overlap with each other. A random subset of 50,000 of the images with labels will be released as validation data included in the development kit along with a list of the 1000 categories. The remaining images will be used for evaluation and will be released without labels at test time.

The training data, the subset of ImageNet containing the 1000 categories and 1.2 million images, will be packaged for easy downloading. The validation and test data for this competition are not contained in the ImageNet training data (we will remove any duplicates).


For each image, algorithms will produce a list of at most 5 object categories in the descending order of confidence. The quality of a labeling will be evaluated based on the label that best matches the ground truth label for the image. The idea is to allow an algorithm to identify multiple objects in an image and not be penalized if one of the objects identified was in fact present, but not included in the ground truth.

There will be two versions of the evaluation criteria: a) non-hierarchical, treating all categories equally, and b) taking into account the hierarchical structure of the set of categories.

For each image, an algorithm will produce 5 labels lj, j=1,...,5. The ground truth labels for the image are gk, k=1,...,n with n objects labeled. The error of the algorithm for that image would be e=1/nΣkminjd(lj,gk). For criteria a)d(x,y)=0 if x=y and 1 otherwise. For criteria b) d(x,y)=height of the lowest common ancestor of x and y in the category hierarchy ( a subset of WordNet ). This is equivalent of predicting a path along the hierarchy and evaluating where the ground truth path and the predicted path diverge. For each criteria the overall error score for an algorithm is the average error over all test images.

Note that for this initial version of the competition, n=1, that is, one ground truth label per image.

Development kit

The development kit will include matlab software to demonstrate training using the ImageNet data (available for download separately from the development kit) and testing on the validation set. This will include routines to compute the overall error score with respect to each criteria.

Timetable (Tentative)


To facilitate easy participation, a set of baseline features will be provided for the images in the 1000 categories in ImageNet and the validation data and later the test data. Routines to demonstrate using this data will be included in the development kit. For 2010, features will include vector quantized SIFT features suitable for a bag of words or spatial pyramid representation.


Test data will be provided in the same format at the validation data, a directory of image files, but not including the labels. Submissions will consist of a text file with one line per image containing identified categories The format is demonstrated in the development kit.


Alex Berg ( Columbia Unviversity ), Jia Deng ( Princeton University ), Fei-Fei Li ( Stanford Unviersity )


Please feel free to send any questions or comments to .