Preprocessing method for DCT-based image-compression

Oizumi, Munenori

doi:10.1109/tce.2006.1706502

Cited by 8 publications

(5 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Variations over this core scheme have been proposed over the years, in which the scale-down or up are optimized for better end-to-end performance. [36], [43], [48], [53] It is well known that compression-decompression often behaves as a denoiser, removing small and faint details from the image. Nevertheless, applying a well-designed denoiser prior to the compression may improve the overall encoding performance by better prioritizing the content to be treated.…”

Section: B Deep-learning Based Solutionsmentioning

confidence: 99%

See 1 more Smart Citation

Better Compression With Deep Pre-Editing

Talebi

Kelly

Luo

et al. 2021

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

Could we compress images via standard codecs while avoiding visible artifacts? The answer is obvious -this is doable as long as the bit budget is generous enough. What if the allocated bit-rate for compression is insufficient? Then unfortunately, artifacts are a fact of life. Many attempts were made over the years to fight this phenomenon, with various degrees of success. In this work we aim to break the unholy connection between bit-rate and image quality, and propose a way to circumvent compression artifacts by pre-editing the incoming image and modifying its content to fit the given bits. We design this editing operation as a learned convolutional neural network, and formulate an optimization problem for its training. Our loss takes into account a proximity between the original image and the edited one, a bit-budget penalty over the proposed image, and a no-reference image quality measure for forcing the outcome to be visually pleasing. The proposed approach is demonstrated on the popular JPEG compression, showing savings in bits and/or improvements in visual quality, obtained with intricate editing effects.

show abstract

Section: B Deep-learning Based Solutionsmentioning

confidence: 99%

“…Far less popular are algorithms that propose to pre-process the image prior to its compression, in order to reduce its entropy, thus avoiding the creation of artifacts in the first place [17], [18], [36], [43], [44], [53]. Indeed, a denoising applied before the compression is often found effective for better encoding performance (e.g.…”

Section: Introductionmentioning

confidence: 99%

Better Compression With Deep Pre-Editing

Talebi

Kelly

Luo

et al. 2021

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

show abstract

“…Pre-Processing Algorithms [33,40,45,50]: It is well known that compression-decompression often behaves as a denoiser, removing small and faint details from the image. Nevertheless, applying a well-designed denoiser prior to the compression may improve the overall encoding performance by better prioritizing the content to be treated.…”

Section: Relation To Prior Workmentioning

confidence: 99%

“…Indeed, hundreds of papers that take this post-processing approach have been published over the years, including recent deep-learning based solutions (e.g., [10,16,20,53]). Far less popular are algorithms that propose to preprocess the image prior to its compression, in order to reduce its entropy, thus avoiding the creation of artifacts in the first place [14,15,33,40,41,50]. Indeed, a denoising applied before the compression is often found effective for better encoding performance (e.g.…”

Section: Introductionmentioning

confidence: 99%

Better Compression with Deep Pre-Editing

Talebi,

Kelly,

Luo

et al. 2020

Preprint

View full text Add to dashboard Cite

Could we compress images via standard codecs while avoiding artifacts? The answer is obvious -this is doable as long as the bit budget is generous enough. What if the allocated bit-rate for compression is insufficient? Then unfortunately, artifacts are a fact of life. Many attempts were made over the years to fight this phenomenon, with various degrees of success. In this work we aim to break the unholy connection between bit-rate and image quality, and propose a way to circumvent compression artifacts by preediting the incoming image and modifying its content to fit the given bits. We design this editing operation as a learned convolutional neural network, and formulate an optimization problem for its training. Our loss takes into account a proximity between the original image and the edited one, a bit-budget penalty over the proposed image, and a noreference image quality measure for forcing the outcome to be visually pleasing. The proposed approach is demonstrated on the popular JPEG compression, showing savings in bits and/or improvements in visual quality, obtained with intricate editing effects.

show abstract

“…The realization of this property in natural images has led to the success of lossy compression standards like JPEG. Even today, the most widely used compression techniques are based on DCT such as JPEG for still images, MP3 for audio and MPEG4 for video [1,2]. One of the reasons behind the success of the DCT based techniques is the availability of fast algorithms to encode the signal [3,4].…”

Section: Introductionmentioning

confidence: 99%

Dictionaries and algorithms for sparsity constrained image reconstruction

Tharmalingam¹

2021

Preprint

View full text Add to dashboard Cite

There has been a growing interest in the different types of dictionaries that can be used in image processing applications. We propose a hybrid dictionary composed of transform based atoms and additional nonlinear atoms generated using the polynomial, rectangular and exponential functions. The additional nonlinear atoms improve signal reconstruction quality for both transient and smooth signals. To further improve signal reconstruction quality, we optimize the hybrid dictionary using training samples from the signal. We also propose a signal coding algorithm that generates additional atoms by performing a circular shift on the provided dictionary prior to coding the signal. We have evaluated the proposed methods against existing predefined dictionaries by visually examining the reconstructed images as well as evaluating the peak signal to noise ratio of the reconstructed signal. All methods proposed in this thesis improved signal reconstruction quality however; we require an in-depth cost analysis study to evaluate its limitations.

show abstract

Preprocessing method for DCT-based image-compression

Cited by 8 publications

References 6 publications

Better Compression With Deep Pre-Editing

Better Compression With Deep Pre-Editing

Better Compression with Deep Pre-Editing

Dictionaries and algorithms for sparsity constrained image reconstruction

Contact Info

Product

Resources

About