ℓ2 optimized predictive image coding with ℓ∞ bound

In many scientific, medical and defense applications of image/video compression, an ℓ_∞ error bound is required. However, pure ℓ_∞-optimized image coding, colloquially known as near-lossless image coding, is prone to structured errors such as contours and speckles if the bit rate is not sufficiently high; moreover, previous ℓ_∞-based image coding methods suffer from poor rate control. In contrast, the ℓ₂ error metric aims for average fidelity and hence preserves the subtlety of smooth waveforms better than the ℓ_∞ error metric and it offers fine granularity in rate control; but pure ℓ₂-based image coding methods (e.g., JPEG 2000) cannot bound individual errors as the ℓ_∞-based methods can. This paper presents a new compression approach to retain the benefits and circumvent the pitfalls of the two error metrics.