Introduction

In this site we publish deep features extracted from various relevant datasets. At the moment the deep features datasets we have published are:

  • YFCC100M-HNfc6
  • 2016Jubilee-HNfc6
  • MicrosoftCOCO-HNfc6
In the following we briefly describe each of them.

All the deep features were extracted using the Caffe framework. In particular we took the activation of the neurons in the fc6 layer of the Hybrid-CNN whose model and weights are public available in the Caffe Model Zoo . The Hybrid-CNN was trained on 1,183 categories (205 scene categories from Places Database and 978 object categories from the train data of ILSVRC2012 (ImageNet) with ~3.6 million images. The architecture is the same as Caffe reference network. More information can be found on the Places-CNN model webapage at MIT .


YFCC100M-HNfc6

The Yahoo Flickr Creative Commons 100M (YFCC100M) dataset was created in 2014 as part of the Yahoo Webscope program. The dataset consists of approximately 99.2 million photos and 0.8 million videos, all uploaded to Flickr between 2004 and 2014 and published under a Creative Commons commercial or non commercial license.

The deep features have been integrated in the corpus maintained by the Multimedia Commons initiative which is an effort to develop and share sets of computed features and ground-truth annotations for the Yahoo Flickr Creative Commons 100 Million dataset (YFCC100M), which contains around 99.2 million images and nearly 800,000 videos from Flickr, all shared under Creative Commons licenses.

We also give Content-Based Image Retrieval results for various approaches and for various subsets of the datasets. While on the web page you can only see 100 results, 10,001 resutls for each query are available for download.

A paper that propose the use of this features and ground-truth results as a benchmark for Similarity Search as been presented at 9th International Conference on Similarity Search and Applications (SISAP):

YFCC100M-HNfc6: A Large-Scale Deep Features Benchmark for Similarity Search
Giuseppe Amato, Fabrizio Falchi, Claudio Gennaro and Fausto Rabitti
International Conference on Similarity Search and Applications. Springer International Publishing, 2016.
Bibtex

A paper that describes the dataset and presents performance indexing has been presented at the Multimedia Commons 2016 Workshop in Amsterdam during ACM Multimedia 2016:

YFCC100M HybridNet fc6 Deep Features for Content-Based Image Retrieval
Giuseppe Amato, Fabrizio Falchi, Claudio Gennaro, and Fausto Rabitti
Proceedings of the 2016 ACM Workshop on Multimedia COMMONS. ACM, 2016
Bibtex


2016Jubilee-HNfc6

About 3.4 million photos related to the 2016 Jubilee of mercy we crawled from Instagram during 5 months. For each of the images, we extracted Deep Features that have been made publicly available together with the URLs of the original Instagram post and image.


MicrosoftCOCO-HNfc6

Microsoft COCO is COCO is an image recognition, segmentation, and captioning dataset. COCO has several features: Object segmentation, Recognition in Context, Multiple objects per image, More than 300,000 images, More than 2 Million instances, 80 object categories, 5 captions per image.
The data has been used in:

Picture It In Your Mind: Generating High Level Visual Representations From Textual Descriptions
Fabio Carrara, Andrea Esuli, Tiziano Fagni, Fabrizio Falchi, Alejandro Moreo Fernández
Bibtex