jhoffman / lsda

Large scale detection through adaptation

Large Scale Detection through Adaptation

Demo code available at: https://github.com/jhoffman/lsda This code accompanies the arXiv report.

Download pre-trained 7.5K model:

• Includes all leaf synsets + 200 detection categories
• Pre-trained Detection Model
• Caffe Weights for 200 strong detectors
• Caffe Weights for 7K weak detectors

To try out an example, run:

startup;
detect10k_demo(rcnn_model, rcnn_feat, 'ex_img.jpg');

Dependency

This code requires Caffe to be installed. For Installation instructions visit: http://caffe.berkeleyvision.org

Citing

"LSDA: Large Scale Detection Through Adaptation." J. Hoffman, 
S. Guadarrama, E. Tzeng, R. Hu, J. Donahue, R. Girshick, T. Darrell, and
K. Saenko. Neural Information Processing Systems (NIPS), 2014.

http://arxiv.org/abs/1407.5035