Energy Compaction-Based Image Compression Using Convolutional AutoEncoder

Cheng, Zhengxue; Sun, Heming; Takeuchi, Masaru; Katto, Jiro

doi:10.1109/tmm.2019.2938345

Cited by 70 publications

(35 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The overhead can be calculated as the third term in the following equation. We compare our results with [27], [18] and [28]. In [27], the kernel size for the main path is 5 × 5 which is same with ours.…”

Section: Memory Consumption Evaluationmentioning

confidence: 83%

“…As a result, at the constraint of the same bit-budget, the QP for large λ is degraded. We also compare with [18] and three traditional standards that are JPEG, JPEG2000 and BPG. For the MS-SSIM, the quantized version can outperform BPG and [18] significantly.…”

Section: B Coding Gain Evaluationmentioning

confidence: 99%

“…We also compare with [18] and three traditional standards that are JPEG, JPEG2000 and BPG. For the MS-SSIM, the quantized version can outperform BPG and [18] significantly. For the PSNR, the quantized version is better than JPEG2000 and [18].…”

Section: B Coding Gain Evaluationmentioning

confidence: 99%

See 2 more Smart Citations

End-To-End Learned Image Compression With Fixed Point Weight Quantization

Sun

Cheng

Takeuchi

et al. 2020

2020 IEEE International Conference on Image Processing (ICIP)

Self Cite

View full text Add to dashboard Cite

End-to-end Learned image compression (LIC) has reached the traditional hand-crafted methods such as BPG (HEVC intra) in terms of the coding gain. However, the large network size prohibits the usage of LIC on resource-limited embedded systems. This paper reduces the network complexity by quantizing both weights and activations. 1) For the weight quantization, we study different kinds of grouping and quantization scheme at first. A channel-wise non-linear quantization scheme is determined based on the coding gain analysis. After that, we propose a fine tuning scheme to clip the weights within a certain range so that the quantization error can be reduced. 2) For the activation quantization, we first propose multiple nonlinear quantization codebooks with different maximum dynamic ranges. By selecting an optimal one through a multiplexer, the quantization range can be saturated to the greatest extent. In addition, we also exploit the mean-removed quantization for the analysis transform outputs in order to reduce the bit-width cost for the specific channel with the large non-zero mean. By quantizing each weight and activation element from 32-bit floating point to 8-bit fixed point, the memory cost for both weight and activation can be reduced by 75% with negligible coding performance loss. As a result, our quantized LIC can still outperform BPG in terms of MS-SSIM. To our best knowledge, this is the first work to give a complete analysis on the coding gain and the memory cost for a quantized LIC network, which validates the feasibility of the hardware implementation.

show abstract

Section: Memory Consumption Evaluationmentioning

confidence: 83%

Section: B Coding Gain Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

End-To-End Learned Image Compression With Fixed Point Weight Quantization

Sun

Cheng

Takeuchi

et al. 2020

2020 IEEE International Conference on Image Processing (ICIP)

Self Cite

View full text Add to dashboard Cite

show abstract

“…This has lead to its incorporation in recent data compression standards and its use in many different cases [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]. In addition, ANS-based encoders could be applicable to a very wide range of multimedia scenarios, such as an alternative to the Rice-Golomb codes employed in the energy-efficient scheme described in [20], as a high-throughput entropy encoder in a high frame rate video format [21], or in general as an entropy encoder in schemes for sparse coding [22], learned image compression [23], compressive sensing [24], or point cloud data compression [25]. In [26], ANS is employed to code large-alphabets sources.…”

Section: Introductionmentioning

confidence: 99%

Redundancy and Optimization of tANS Entropy Encoders

Blanes¹,

Hernández-Cabronero²,

Serra-Sagristà³

et al. 2021

IEEE Trans. Multimedia

View full text Add to dashboard Cite

Nowadays entropy encoders are part of almost all data compression methods, with the Asymmetrical Numeral Systems (ANS) family of entropy encoders having recently risen in popularity. Entropy encoders based on the tabled variant of ANS are known to provide varying performances depending on their internal design. In this paper, we present a method that calculates encoder redundancies in almost linear time, which translates in practice to thousand-fold speedups in redundancy calculations for small automatons, and allows redundancy calculations for automatons with tens of millions of states that would be otherwise prohibitive. We also address the problem of improving tabled ANS encoder designs, by employing the aforementioned redundancy calculation method in conjunction with a stochastic hill climbing strategy. The proposed approach consistently outperforms state-of-the-art methods in tabled ANS encoder design. For automatons of twice the alphabet size, experimental results show redundancy reductions around 10% over the default initialization method and over 30% for random initialization.

show abstract

“…In [21, 22], the authors discussed the usage of ANN for image compression. The applicability of the CNN for image compression is explored in [23–26]. Furthermore, methods that utilise sparse representation based on dictionary learning algorithms have proven to be effective techniques for image compression applications [27–29].…”

Section: Introductionmentioning

confidence: 99%

Image compression using adaptive multiresolution image decomposition algorithm

Alkishriwo

2020

IET image process

View full text Add to dashboard Cite

Energy Compaction-Based Image Compression Using Convolutional AutoEncoder

Cited by 70 publications

References 20 publications

End-To-End Learned Image Compression With Fixed Point Weight Quantization

End-To-End Learned Image Compression With Fixed Point Weight Quantization

Redundancy and Optimization of tANS Entropy Encoders

Image compression using adaptive multiresolution image decomposition algorithm

Contact Info

Product

Resources

About