* rewritten the following layers to be [much] more efficient: convolution, fully connected, activations (relu, tanh, ...), LRN. Use optional AVX optimization for the first two.

* eliminated trailing whitespaces

The while loop in member function buildPyramid in deepflow.cpp terminated as a result of the break statement and never as a result of the termination criteria because the "i" variable in the while loop was never incremented.

Changed the while loop into a for loop

Changed the while loop into a for loop in the member function buildPyramid in deepFlow.cpp

removed trailing whitespace

* Reuse deep learning output blobs

* Changed order for iterating through blobs while seeking memory. Refactored a little.

fix java wrappers for ERFilter, OCRHMMDecoder, OCRBeamSearchDecoder (#1210)

* fix java wrappers for ERFilter, OCRHMMDecoder, OCRBeamSearchDecoder

* fix comments

Fixed a buffer overflow which occurs when setImage is called with different images and default image_id.

Incremental commit.

Theory updated.

Theory finished.

Inpainting in progress.

Inpainting done plus minor fix in kernel computation.

Tutorials done.

Documentation done.

Whitespace fix.

Update.

Conversion to jpg.

Documentation links fixed.

Made separate functions for computing output shapes for all layers. Removed output blobs allocation from layers