DB2DPCA

MATLAB experiments for bilateral 2DPCA information fusion in image reconstruction and recognition.

Paper

Scripts for Fusion of Bilateral 2DPCA Information for Image Reconstruction and Recognition. Read the paper.

Prerequisites and Execution

The datasets are external to this repository. Both classification.m and reconstruction.m default to ../data/handwritten/BDL; database.m lists other example locations.

Prepare the data expected by the selected classify_*.m or reco_*.m function, then edit the database variable to the actual path. Classification defaults to classfier='CRC' and nPC=16. From the repository directory in MATLAB, run the required stage:

classification
% Or, for reconstruction:
reconstruction

The current checkout does not include the default BDL dataset or a dataset-conversion script.

Repository Structure

• classification.m and reconstruction.m: experiment entry points.
• database.m: example external input locations.
• classify_DB2DPCA.m and reco_DB2DPCA.m: DB2DPCA experiment functions.
• CRC.m: CRC classifier.

License

See the existing GPL-3.0 license.