Image compression

1. predict the value of the next pixel

take the diff from the actual value. If your predictor did a decent job, the diff should be close to zero for many pixels

a typical (and fairly good) predcitor predicts the next pixels has the same value as the previous one

2. then use entropy coders

• huffman encoder (easy to employee, but suboptimal)
• range encoder (I used this at least three times, based one okanohara-san's(!) implementation)
• arithmetic encoders