Skip to main content


Fig. 2 | EURASIP Journal on Advances in Signal Processing

Fig. 2

From: Fast dictionary learning from incomplete data

Fig. 2

Recovery performance of the corruption adapted versus the unadapted learning algorithms for the DCT (left) and the random (right) representation pair. The recovery performance is measured in terms of the low-rank recovery error (top) and the percentage of unrecovered dictionary atoms from a random initialisation (bottom)

Back to article page