QPAC: A Novel Document Image Compression Technique Based on ANFIS for Calibrated Quality Preservation

摘要

In the present paper, a novel compression technique is used to compress the document images without compromising abrupt quality degradation. An Adaptive neuro-fuzzy inference system (ANFIS) based classification scheme is used for segmenting the input document followed by some neighborhood smoothing is performed. A tuning based adaptive compression scheme is used for compressing the ANFIS classified data. A 3D trade-off between quality, compression ratio and relative occupancy of image region is addressed in a calibrated manner.