JPEG2000 code-stream interpreter

Abstract

A code-stream interpreter able to parse and determine image features, file attributes, and perform error detection utilizing an effective error classification method for JPEG2000 compressed images is proposed. The proposed implementation performs analysis on a JPEG2000 image to provide insight primarily for JPEG2000 encoder and decoder research and development. Results of the interpreter presented in this paper are demonstrated for various JPEG2000 compressed images. Parameter changes are made on these test images to both emulate different types of encoders and simulate prototype encoder bugs, even simulating a small amount of data loss which can occur after going through a transmission channel. The resulting JPEG2000 code-streams are then decoded by mainstream, commercial image decoders with JPEG2000 decoding capabilities and also tested with the proposed JPEG2000 code-stream interpreter. Results are demonstrated which reveal that the proposed interpreter provides information about the image which is occasionally unknown, overlooked, or even sometimes misinterpreted by commercial image decoders. Due to its dynamic error detection and classification capabilities, the JPEG2000 code-stream interpreter has potential to be adapted as an automated error correction system as a future area of research for either pre-decoder processing, or post-encoder processing after data transmission losses.