In a significant advancement for digital imaging, Professor Marko Huhtanen from the University of Oulu has introduced a groundbreaking method for compressing images. This innovative technique, detailed in the latest edition of IEEE Signal Processing Letters, combines several established compression methods to enhance efficiency and flexibility.
Digital photography predominantly relies on the JPEG format, known for its widespread use for both casual and professional purposes. While many photographers favor RAW format due to its adaptability for post-processing, the JPEG format retains only between 10% and 25% of the original image data. The significance of this data loss varies among viewers, posing a common challenge for anyone who captures or shares digital images.
The process of converting images into a transmittable format is straightforward. As Professor Huhtanen explains, “We don’t see a perfect image because the amount of information is infinite. So we must compress and retain only the essential, sufficient data. This is done mathematically in a way that must also be algorithmically fast.”
Revolutionizing Compression Techniques
Huhtanen’s method innovatively processes images both horizontally and vertically by utilizing diagonal matrices, constructing image approximations layer by layer. This technique mirrors a simplified version of Berlekamp’s switching game, adapted for continuous applications.
“Image compression is a fundamental problem in imaging—how to pack an image into the smallest possible space for fast transmission and sharing. The original image takes up too much space in computer memory, so we aim to preserve only 10% to 25% of the image’s information,” Huhtanen noted.
Current technologies such as JPEG rely on algorithms developed almost half a century ago by Nazir Ahmed, an American professor of electrical and computer engineering. Although Ahmed initially sought to base his compression on Principal Component Analysis (PCA), he faced algorithmic challenges that led to the creation of a simpler method utilizing the Discrete Cosine Transform (DCT). Despite initial funding rejections due to perceived simplicity, Ahmed’s DCT method became a standard in image compression.
“Scientific publishing involves a lot of randomness, and it is hard to predict what will ultimately be considered significant. In this case, significance is also relative,” Huhtanen commented on the evolution of these techniques.
Integrating DCT and PCA for Better Results
The primary goal of any compression technique is to eliminate as much data as possible without noticeable differences in the resulting image. Huhtanen compares JPEG’s method to a straightforward approach, where each image is divided into 64 segments, each compressed via DCT. While effective in practice, he notes the method is not particularly complex mathematically.
PCA, which Ahmed initially proposed, was ultimately overlooked in imaging due to its perceived labor intensity and rigidity. Huhtanen’s research successfully merges these two approaches, enabling a more flexible application of both DCT and PCA. “I managed to remove this rigidity, allowing the ideas to be mixed and the best aspects of both to be utilized,” he explained.
Although he refrains from predicting the future applicability of his findings, Huhtanen asserts that he has addressed a long-standing issue within the field of image compression. A broad range of algorithms has emerged from his research, with PCA serving as just one special case.
Understanding PCA can be illustrated through a comparison to film photography. Digital compression can be likened to transforming an image into a “negative,” from which necessary elements are extracted to create a visible image. The recipient ultimately receives this “negative form,” which is then rendered into a format that can be viewed.
The implications of Huhtanen’s method extend beyond just improved image quality. Enhanced compression techniques can lead to faster transmission speeds, reduced storage space, and lower energy consumption. “Individual components arrive through the channel, and the image sharpens as the compression is decompressed. If this can be done better than it is currently, image transfer speeds up and more information can be transmitted,” he said.
By reducing data volume smartly, Huhtanen’s approach not only optimizes storage and speeds up transmission but also facilitates faster computation processes, making it well-suited for parallel data processing. The layered construction of images allows for greater control and adjustment during compression, ultimately leading to more efficient energy use.
As the demand for high-quality digital images continues to grow, Huhtanen’s innovative compression method may well represent a pivotal advancement in how images are stored and transmitted across the globe.
