To be clear, you're singling out a specific algorithm and offering a 1000USD, even money bet that it will be used by the winner?

yep

The "state of the art" they reference is SVM's trained on color and texture features.

Pre deep belief network I'd agree with your guess on convolutional neural networks. However, now I'd guess you'd use a deep belief network to create a network that would pick out better features than those picked out "by hand" in the convolutional neural network. (See for example [1][2])

So my money would be on some deep belief network.

[1] Hinton, G. E, Osindero, S., and Teh, Y. W. (2006). A fast learning algorithm for deep belief nets. Neural Computation, 18:1527-1554.

[2] Building high-level features using large scale unsupervised learning arXiv:1112.6209

So far as it comes to large datasets unsupervised learning doesn't work ! You better off training initially discriminatively your network on imagenet, and then switch to this cat vs dog training. Rather, than do unsupervised learning.

program it in Lush then.

everyone here found out about Deep Neural Networks and that is all they know.

Whole bunch of stuff on RBM and Deep belief nets. Also, has results on a competition on recognizing 1000 objects.

http://www.cs.toronto.edu/~hinton/

Yes, it'd be surprised if the straightforward implementation from https://code.google.com/p/cuda-convnet/, run on a GPU with lots of transformations, wasn't the winning entry.