arXiv subjects

Ezgi Ozyilkan

Publications and source records attributed to Ezgi Ozyilkan.

15 recordsLinked to original sources

arXiv2026

Drop-In Perceptual Optimization for 3D Gaussian Splatting

Despite their output being ultimately consumed by human viewers, 3D Gaussian Splatting (3DGS) methods often rely on ad-hoc combinations of pixel-level losses, resulting in blurry renderings. To address this, we systematically explore perceptual optimization strategies for 3DGS by searching over a diverse set of distortion losses. We conduct the first-of-its-kind large-scale human subjective study on 3DGS, involving 39,320 pairwise ratings across several datasets and 3DGS frameworks. A regularized version of Wasserstein Distortion, which we call WD-R, emerges as the clear winner, excelling at recovering fine textures without incurring a higher splat count. WD-R is preferred by raters more than $2.3\times$ over the original 3DGS loss, and $1.5\times$ over the current best method Perceptual-GS. WD-R also consistently achieves state-of-the-art LPIPS, DISTS, and FID scores across various datasets, and generalizes across recent frameworks, such as Mip-Splatting and Scaffold-GS, where replacing the original loss with WD-R consistently enhances perceptual quality within a similar resource budget (number of splats for Mip-Splatting, model size for Scaffold-GS), and leads to reconstructions being preferred by human raters $1.8\times$ and $3.6\times$, respectively. We also find that this carries over to the task of 3DGS scene compression, with $\approx 50\%$ bitrate savings for comparable perceptual metric performance.

cs.CV↗

arXiv2026

SoftBinary Coding: A New Information-Theoretic Neural Compression Paradigm

Neural compression is currently dominated by Nonlinear Transform Coding (NTC), which maps data to real-valued latents via continuous transforms. Despite its success, NTC suffers from train-test mismatch due to non-differentiable quantization, a ``smoothness bias" inherent in continuous transforms that precludes optimality for certain sources, and a loss of ``shaping gain" due to the complexity of including high-dimensional vector quantization. We propose SoftBinary Coding (SBC), an end-to-end learning paradigm that bypasses these limitations by using a stochastic binary latent space. In the spirit of vector quantization, SBC employs discrete representations and compresses them through a novel fast binary channel simulation scheme, for which we provide a proof of rate optimality. Experimental gains on information-theoretic sources provide both theoretical and practical closure to NTC's limitations, establishing discrete binary structures as a viable path toward reaching optimal rate--distortion bounds. Surprisingly, SBC also achieves state-of-the-art performance on vector quantization of i.i.d. sources, exceeding Trellis Coded Quantization of the Gaussian source.

cs.IT↗

arXiv2025

Neural Compress-and-Forward for the Primitive Diamond Relay Channel

The diamond relay channel, where a source communicates with a destination via two parallel relays, is one of the canonical models for cooperative communications. We focus on the primitive variant, where each relay observes a noisy version of the source signal and forwards a compressed description over an orthogonal, noiseless, finite-rate link to the destination. Compress-and-forward (CF) is particularly effective in this setting, especially under oblivious relaying where relays lack access to the source codebook. While neural CF methods have been studied in single-relay channels, extending them to the two-relay case is non-trivial, as it requires fully distributed compression without any inter-relay coordination. We demonstrate that learning-based quantizers at the relays can harness input correlations by operating remote, yet in a collaborative fashion, enabling effective distributed compression in line with Berger-Tung-style coding. Each relay separately compresses its observation using a one-shot learned quantizer, and the destination jointly decodes the source message. Simulation results show that the proposed scheme, trained end-to-end with finite-order modulation, operates close to the known theoretical bounds. These results demonstrate that neural CF can scale to multi-relay systems while maintaining both performance and interpretability.

cs.IT↗

arXiv2025

Learning to Write on Dirty Paper

Dirty paper coding (DPC) is a classical problem in information theory that considers communication in the presence of channel state known only at the transmitter. While the theoretical impact of DPC has been substantial, practical realizations of DPC, such as Tomlinson-Harashima precoding (THP) or lattice-based schemes, often rely on specific modeling assumptions about the input, state and channel. In this work, we explore whether modern learning-based approaches can offer a complementary path forward by revisiting the DPC problem. We propose a data-driven solution in which both the encoder and decoder are parameterized by neural networks. Our proposed model operates without prior knowledge of the state (also referred to as "interference"), channel or input statistics, and recovers nonlinear mappings that yield effective interference pre-cancellation. To the best of our knowledge, this is the first interpretable proof-of-concept demonstrating that learning-based DPC schemes can recover characteristic features of well-established solutions, such as THP and lattice-based precoding, and outperform them in several regimes.

cs.IT↗

arXiv2025

Learning-Based Compress-and-Forward Schemes for the Relay Channel

The relay channel, consisting of a source-destination pair along with a relay, is a fundamental component of cooperative communications. While the capacity of a general relay channel remains unknown, various relaying strategies, including compress-and-forward (CF), have been proposed. In CF, the relay forwards a quantized version of its received signal to the destination. Given the correlated signals at the relay and destination, distributed compression techniques, such as Wyner--Ziv coding, can be harnessed to utilize the relay-to-destination link more efficiently. Leveraging recent advances in neural network-based distributed compression, we revisit the relay channel problem and integrate a learned task-aware Wyner--Ziv compressor into a primitive relay channel with a finite-capacity out-of-band relay-to-destination link. The resulting neural CF scheme demonstrates that our compressor recovers binning of the quantized indices at the relay, mimicking the optimal asymptotic CF strategy, although no structure exploiting the knowledge of source statistics was imposed into the design. The proposed neural CF, employing finite order modulation, operates closely to the rate achievable in a primitive relay channel with a Gaussian codebook. We showcase the advantages of exploiting the correlated destination signal for relay compression through various neural CF architectures that involve end-to-end training of the compressor and the demodulator components. Our learned task-oriented compressors provide the first proof-of-concept work toward interpretable and practical neural CF relaying schemes.

cs.IT↗

arXiv2024

Robust Distributed Compression with Learned Heegard-Berger Scheme

We consider lossy compression of an information source when decoder-only side information may be absent. This setup, also referred to as the Heegard-Berger or Kaspi problem, is a special case of robust distributed source coding. Building upon previous works on neural network-based distributed compressors developed for the decoder-only side information (Wyner-Ziv) case, we propose learning-based schemes that are amenable to the availability of side information. We find that our learned compressors mimic the achievability part of the Heegard-Berger theorem and yield interpretable results operating close to information-theoretic bounds. Depending on the availability of the side information, our neural compressors recover characteristics of the point-to-point (i.e., with no side information) and the Wyner-Ziv coding strategies that include binning in the source space, although no structure exploiting knowledge of the source and side information was imposed into the design.

cs.IT↗

arXiv2024

Neural Compress-and-Forward for the Relay Channel

The relay channel, consisting of a source-destination pair and a relay, is a fundamental component of cooperative communications. While the capacity of a general relay channel remains unknown, various relaying strategies, including compress-and-forward (CF), have been proposed. For CF, given the correlated signals at the relay and destination, distributed compression techniques, such as Wyner-Ziv coding, can be harnessed to utilize the relay-to-destination link more efficiently. In light of the recent advancements in neural network-based distributed compression, we revisit the relay channel problem, where we integrate a learned one-shot Wyner--Ziv compressor into a primitive relay channel with a finite-capacity and orthogonal (or out-of-band) relay-to-destination link. The resulting neural CF scheme demonstrates that our task-oriented compressor recovers "binning" of the quantized indices at the relay, mimicking the optimal asymptotic CF strategy, although no structure exploiting the knowledge of source statistics was imposed into the design. We show that the proposed neural CF scheme, employing finite order modulation, operates closely to the capacity of a primitive relay channel that assumes a Gaussian codebook. Our learned compressor provides the first proof-of-concept work toward a practical neural CF relaying scheme.

cs.IT↗

arXiv2024

Distributed Deep Joint Source-Channel Coding with Decoder-Only Side Information

We consider low-latency image transmission over a noisy wireless channel when correlated side information is present only at the receiver side (the Wyner-Ziv scenario). In particular, we are interested in developing practical schemes using a data-driven joint source-channel coding (JSCC) approach, which has been previously shown to outperform conventional separation-based approaches in the practical finite blocklength regimes, and to provide graceful degradation with channel quality. We propose a novel neural network architecture that incorporates the decoder-only side information at multiple stages at the receiver side. Our results demonstrate that the proposed method succeeds in integrating the side information, yielding improved performance at all channel conditions in terms of the various quality measures considered here, especially at low channel signal-to-noise ratios (SNRs) and small bandwidth ratios (BRs). We have made the source code of the proposed method public to enable further research, and the reproducibility of the results.

cs.CV↗

arXiv2024

Distributed Compression in the Era of Machine Learning: A Review of Recent Advances

Many applications from camera arrays to sensor networks require efficient compression and processing of correlated data, which in general is collected in a distributed fashion. While information-theoretic foundations of distributed compression are well investigated, the impact of theory in practice-oriented applications to this day has been somewhat limited. As the field of data compression is undergoing a transformation with the emergence of learning-based techniques, machine learning is becoming an important tool to reap the long-promised benefits of distributed compression. In this paper, we review the recent contributions in the broad area of learned distributed compression techniques for abstract sources and images. In particular, we discuss approaches that provide interpretable results operating close to information-theoretic bounds. We also highlight unresolved research challenges, aiming to inspire fresh interest and advancements in the field of learned distributed compression.

cs.IT↗

arXiv2023

Neural Distributed Compressor Discovers Binning

We consider lossy compression of an information source when the decoder has lossless access to a correlated one. This setup, also known as the Wyner-Ziv problem, is a special case of distributed source coding. To this day, practical approaches for the Wyner-Ziv problem have neither been fully developed nor heavily investigated. We propose a data-driven method based on machine learning that leverages the universal function approximation capability of artificial neural networks. We find that our neural network-based compression scheme, based on variational vector quantization, recovers some principles of the optimum theoretical solution of the Wyner-Ziv setup, such as binning in the source space as well as optimal combination of the quantization index and side information, for exemplary sources. These behaviors emerge although no structure exploiting knowledge of the source distributions was imposed. Binning is a widely used tool in information theoretic proofs and methods, and to our knowledge, this is the first time it has been explicitly observed to emerge from data-driven learning.

cs.IT↗

arXiv2023

Learned Wyner-Ziv Compressors Recover Binning

We consider lossy compression of an information source when the decoder has lossless access to a correlated one. This setup, also known as the Wyner-Ziv problem, is a special case of distributed source coding. To this day, real-world applications of this problem have neither been fully developed nor heavily investigated. We propose a data-driven method based on machine learning that leverages the universal function approximation capability of artificial neural networks. We find that our neural network-based compression scheme re-discovers some principles of the optimum theoretical solution of the Wyner-Ziv setup, such as binning in the source space as well as linear decoder behavior within each quantization index, for the quadratic-Gaussian case. These behaviors emerge although no structure exploiting knowledge of the source distributions was imposed. Binning is a widely used tool in information theoretic proofs and methods, and to our knowledge, this is the first time it has been explicitly observed to emerge from data-driven learning.

cs.IT↗

arXiv2023

Learned Disentangled Latent Representations for Scalable Image Coding for Humans and Machines

As an increasing amount of image and video content will be analyzed by machines, there is demand for a new codec paradigm that is capable of compressing visual input primarily for the purpose of computer vision inference, while secondarily supporting input reconstruction. In this work, we propose a learned compression architecture that can be used to build such a codec. We introduce a novel variational formulation that explicitly takes feature data relevant to the desired inference task as input at the encoder side. As such, our learned scalable image codec encodes and transmits two disentangled latent representations for object detection and input reconstruction. We note that compared to relevant benchmarks, our proposed scheme yields a more compact latent representation that is specialized for the inference task. Our experiments show that our proposed system achieves a bit rate savings of 40.6% on the primary object detection task compared to the current state-of-the-art, albeit with some degradation in performance for the secondary input reconstruction task.

eess.IV↗

arXiv2023

Neural Distributed Image Compression with Cross-Attention Feature Alignment

We consider the problem of compressing an information source when a correlated one is available as side information only at the decoder side, which is a special case of the distributed source coding problem in information theory. In particular, we consider a pair of stereo images, which have overlapping fields of view, and are captured by a synchronized and calibrated pair of cameras as correlated image sources. In previously proposed methods, the encoder transforms the input image to a latent representation using a deep neural network, and compresses the quantized latent representation losslessly using entropy coding. The decoder decodes the entropy-coded quantized latent representation, and reconstructs the input image using this representation and the available side information. In the proposed method, the decoder employs a cross-attention module to align the feature maps obtained from the received latent representation of the input image and a latent representation of the side information. We argue that aligning the correlated patches in the feature maps allows better utilization of the side information. We empirically demonstrate the competitiveness of the proposed algorithm on KITTI and Cityscape datasets of stereo image pairs. Our experimental results show that the proposed architecture is able to exploit the decoder-only side information in a more efficient manner compared to previous works.

eess.IV↗

arXiv2021

Neural Distributed Image Compression using Common Information

We present a novel deep neural network (DNN) architecture for compressing an image when a correlated image is available as side information only at the decoder. This problem is known as distributed source coding (DSC) in information theory. In particular, we consider a pair of stereo images, which generally have high correlation with each other due to overlapping fields of view, and assume that one image of the pair is to be compressed and transmitted, while the other image is available only at the decoder. In the proposed architecture, the encoder maps the input image to a latent space, quantizes the latent representation, and compresses it using entropy coding. The decoder is trained to extract the common information between the input image and the correlated image, using only the latter. The received latent representation and the locally generated common information are passed through a decoder network to obtain an enhanced reconstruction of the input image. The common information provides a succinct representation of the relevant information at the receiver. We train and demonstrate the effectiveness of the proposed approach on the KITTI and Cityscape datasets of stereo image pairs. Our results show that the proposed architecture is capable of exploiting the decoder-only side information, and outperforms previous work on stereo image compression with decoder side information.

eess.IV↗

arXiv2020

Deep Joint Transmission-Recognition for Multi-View Cameras

We propose joint transmission-recognition schemes for efficient inference at the wireless edge. Motivated by the surveillance applications with wireless cameras, we consider the person classification task over a wireless channel carried out by multi-view cameras operating as edge devices. We introduce deep neural network (DNN) based compression schemes which incorporate digital (separate) transmission and joint source-channel coding (JSCC) methods. We evaluate the proposed device-edge communication schemes under different channel SNRs, bandwidth and power constraints. We show that the JSCC schemes not only improve the end-to-end accuracy but also simplify the encoding process and provide graceful degradation with channel quality.

cs.LG↗