Jpeg coder with limited file size, its method and recording medium

TLDR: In this paper, the problem of reducing a compressed file size furthermore by deciding a cut-off number in a 1st array with coefficients sequenced by frequencies and nullifying frequency coefficients of the first array whose numbers are over the cutoff number and whose values are less than a corresponding threshold is solved.

Abstract: PROBLEM TO BE SOLVED: To reduce a compressed file size furthermore by deciding a cut-off number in a 1st array with coefficients sequenced by frequencies and nullifying frequency coefficients of the 1st array whose numbers are over the cut-off number and whose values are less than a corresponding threshold. SOLUTION: A compressed bit stream of a current JPEG file 17a is fed to a Huffman decoder 19, from which 8×8 blocks consisting of quantized DCT coefficients are reproduced. The coefficient blocks are given to a JPEG coder (FSBJT) 31 with a limited file size, where they are encoded. In this case, a 1st path is generated for data, and when a cut-off number is defined to be, e.g. 61, all of 61st, 62nd and 63rd coefficients are compared with a threshold value to generate an array whose element values can be saved for the compression purpose. After DCT data are coded, the result is given to a Huffman encoder 16, from which a new JPEG file 17b is generated.