No icon

Camera-Aware Multi-Resolution Analysis for Raw Image Sensor Data Compression

Camera-Aware Multi-Resolution Analysis for Raw Image Sensor Data Compression

Abstract:

We propose novel lossless and lossy compression schemes for color filter array (CFA) sampled images based on the Camera-Aware Multi-Resolution Analysis, or CAMRA. Specifically, by CAMRA we refer to modifications that we make to wavelet transform of CFA sampled images in order to achieve a very high degree of decorrelation at the finest scale wavelet coefficients; and a series of color processing steps applied to the coarse scale wavelet coefficients, aimed at limiting the propagation of lossy compression errors through the subsequent camera processing pipeline. We validated our theoretical analysis and the performance of the proposed compression schemes using the images of natural scenes captured in a raw format. The experimental results verify that our proposed methods improve coding efficiency relative to the standard and the state-of-the-art compression schemes for CFA sampled images.

Existing System:

When compressing the image data, a digital data representation of the best quality is highly desirable in many applications, such as digital cinema, broadcast, medical imaging, image archives, etc. In such applications, lossless compression is preferable so that the original data can be reconstructed perfectly. However, lossless compression often requires large storage space. Hence, lossy compression, which could significantly reduce data size for storing, processing, and transmitting, is more common in consumer applications. On the other hand, photographers often work directly with raw sensor data to maximize control over the post-processing.

Proposed System:

We propose novel lossless and lossy compression schemes for raw sensor data based on Camera-Aware Multi-Resolution Analysis, or CAMRA. The wavelet analysis in CAMRA is “camera-aware” in the sense that it explicitly handles the correlation in the finest scale wavelet coefficients caused by CFA sampling, and it models the camera processing pipeline within the wavelet domain. Our work is inspired by the CFA compression method, the wavelet sampling theory, and the demosaicking work. Specifically, the CFA sampled image compression scheme leveraged a heuristic observation made about the behavior of the wavelet transform coefficients corresponding to the CFA sampled image. A more rigorous analysis of wavelet sampling revealed that fine scale wavelet coefficients are highly redundant, and the coarse scale wavelet coefficients have direct correspondences to the sensor color spaces. We previously used these facts to develop demosaicking and denoising methods. In this paper, the analysis was leveraged to decorrelate the redundant coefficients in the finest scale wavelet and to apply color processing to the coarse scale wavelets coefficients in order to improve the compression performance.

Comment As:

Comment (0)