Useful Resources

If you find additional resources, email me a link and I will add them to this page.

Data and downloading scripts

Flickr - code to download images and meta-data via text queries available here.

MIRFlickr Collection - 1 million photos with tags, available here.

SBU Captioned Photo Dataset - 1 million photos with captions, available here. Pre-processed object detections and scene classifications available here

ImageNet - 14 million labeled images for nouns, some synsets with bounding boxes or attributes, available here.

LabelMe - Images labeled with polygons and categories for depicted objects, available here.

General visual representation code

Gist feature code - available here.

SIFT/HOG/MSER feature code and assorted algorithms - available here.

Useful Natural Language Processing code

Berkeley parser - code available here.

Part of speech tagger - code available here.

Wordnet - code available here.

Code for Object Detection

Deformable Part Models for detection - code available here.

Code specific to faces

Face detection and feature representation - code available here.

Face attribute prediction service - available here.