Search

We propose to replace vector quantization (VQ) in the latent representation of VQ-VAEs with a simple scheme termed finite scalar quantization (FSQ), where we project the VAE representation down to a few dimensions (typically less than 10). Each dimension is quantized to a small set of fixed values, leading to an (implicit) codebook given by the product of these sets. By appropriately choosing the number of dimensions and values each dimension can take, we obtain the same codebook size as in VQ. On top of such discrete representations, we can train the same models that have been trained on VQ-VAE representations. For example, autoregressive and masked transformer models for image generation, multimodal generation, and dense prediction computer vision tasks. Concretely, we employ FSQ with MaskGIT for image generation, and with UViM for depth estimation, colorization, and panoptic segmentation. Despite the much simpler design of FSQ, we obtain competitive performance in all these tasks. We emphasize that FSQ does not suffer from codebook collapse and does not need the complex machinery employed in VQ (commitment losses, codebook reseeding, code splitting, entropy penalties, etc.) to learn expressive discrete representations., Comment: Code: https://github.com/google-research/google-research/tree/master/fsq

Despite the tremendous success of diffusion generative models in text-to-image generation, replicating this success in the domain of image compression has proven difficult. In this paper, we demonstrate that diffusion can significantly improve perceptual quality at a given bit-rate, outperforming state-of-the-art approaches PO-ELIC and HiFiC as measured by FID score. This is achieved using a simple but theoretically motivated two-stage approach combining an autoencoder targeting MSE followed by a further score-based decoder. However, as we will show, implementation details matter and the optimal design decisions can differ greatly from typical text-to-image models.

We show how bidirectional transformers trained for masked token prediction can be applied to neural image compression to achieve state-of-the-art results. Such models were previously used for image generation by progressivly sampling groups of masked tokens according to uncertainty-adaptive schedules. Unlike these works, we demonstrate that predefined, deterministic schedules perform as well or better for image compression. This insight allows us to use masked attention during training in addition to masked inputs, and activation caching during inference, to significantly speed up our models (~4 higher inference speed) at a small increase in bitrate.

By optimizing the rate-distortion-realism trade-off, generative compression approaches produce detailed, realistic images, even at low bit rates, instead of the blurry reconstructions produced by rate-distortion optimized models. However, previous methods do not explicitly control how much detail is synthesized, which results in a common criticism of these methods: users might be worried that a misleading reconstruction far from the input image is generated. In this work, we alleviate these concerns by training a decoder that can bridge the two regimes and navigate the distortion-realism trade-off. From a single compressed representation, the receiver can decide to either reconstruct a low mean squared error reconstruction that is close to the input, a realistic reconstruction with high perceptual quality, or anything in between. With our method, we set a new state-of-the-art in distortion-realism, pushing the frontier of achievable distortion-realism pairs, i.e., our method achieves better distortions at high realism and better realism at low distortion than ever before., Comment: CVPR'23 Camera Ready

We show how transformers can be used to vastly simplify neural video compression. Previous methods have been relying on an increasing number of architectural biases and priors, including motion prediction and warping operations, resulting in complex models. Instead, we independently map input frames to representations and use a transformer to model their dependencies, letting it predict the distribution of future representations given the past. The resulting video compression transformer outperforms previous methods on standard video compression data sets. Experiments on synthetic data show that our model learns to handle complex motion patterns such as panning, blurring and fading purely from data. Our approach is easy to implement, and we release code to facilitate future research., Comment: NeurIPS'22 Camera Ready Version. Code: https://goo.gle/vct-paper

We present the first neural video compression method based on generative adversarial networks (GANs). Our approach significantly outperforms previous neural and non-neural video compression methods in a user study, setting a new state-of-the-art in visual quality for neural methods. We show that the GAN loss is crucial to obtain this high visual quality. Two components make the GAN loss effective: we i) synthesize detail by conditioning the generator on a latent extracted from the warped previous reconstruction to then ii) propagate this detail with high-quality flow. We find that user studies are required to compare methods, i.e., none of our quantitative metrics were able to predict all studies. We present the network design choices in detail, and ablate them with user studies., Comment: First two authors contributed equally. ECCV Camera ready version

Stochastic encoders have been used in rate-distortion theory and neural compression because they can be easier to handle. However, in performance comparisons with deterministic encoders they often do worse, suggesting that noise in the encoding process may generally be a bad idea. It is poorly understood if and when stochastic encoders do better than deterministic encoders. In this paper we provide one illustrative example which shows that stochastic encoders can significantly outperform the best deterministic encoders. Our toy example suggests that stochastic encoders may be particularly useful in the regime of "perfect perceptual quality".

We review a class of methods that can be collected under the name nonlinear transform coding (NTC), which over the past few years have become competitive with the best linear transform codecs for images, and have superseded them in terms of rate--distortion performance under established perceptual quality metrics such as MS-SSIM. We assess the empirical rate--distortion performance of NTC with the help of simple example sources, for which the optimal performance of a vector quantizer is easier to estimate than with natural data sources. To this end, we introduce a novel variant of entropy-constrained vector quantization. We provide an analysis of various forms of stochastic optimization techniques for NTC models; review architectures of transforms based on artificial neural networks, as well as learned entropy models; and provide a direct comparison of a number of methods to parameterize the rate--distortion trade-off of nonlinear transforms, introducing a simplified one., Comment: 17 pages, 14 figures. Accepted for publication in IEEE Journal of Selected Topics in Signal Processing

We extensively study how to combine Generative Adversarial Networks and learned compression to obtain a state-of-the-art generative lossy compression system. In particular, we investigate normalization layers, generator and discriminator architectures, training strategies, as well as perceptual losses. In contrast to previous work, i) we obtain visually pleasing reconstructions that are perceptually similar to the input, ii) we operate in a broad range of bitrates, and iii) our approach can be applied to high-resolution images. We bridge the gap between rate-distortion-perception theory and practice by evaluating our approach both quantitatively with various perceptual metrics, and with a user study. The study shows that our method is preferred to previous approaches even if they use more than 2x the bitrate., Comment: This is the Camera Ready version for NeurIPS 2020. Project page: https://hific.github.io

A popular approach to learning encoders for lossy compression is to use additive uniform noise during training as a differentiable approximation to test-time quantization. We demonstrate that a uniform noise channel can also be implemented at test time using universal quantization (Ziv, 1985). This allows us to eliminate the mismatch between training and test phases while maintaining a completely differentiable loss function. Implementing the uniform noise channel is a special case of the more general problem of communicating a sample, which we prove is computationally hard if we do not make assumptions about its distribution. However, the uniform special case is efficient as well as easy to implement and thus of great interest from a practical point of view. Finally, we show that quantization can be obtained as a limiting case of a soft quantizer applied to the uniform noise channel, bridging compression with and without quantization., Comment: Authors contributed equally

We address interactive full image annotation, where the goal is to accurately segment all object and stuff regions in an image. We propose an interactive, scribble-based annotation framework which operates on the whole image to produce segmentations for all regions. This enables sharing scribble corrections across regions, and allows the annotator to focus on the largest errors made by the machine across the whole image. To realize this, we adapt Mask-RCNN into a fast interactive segmentation framework and introduce an instance-aware loss measured at the pixel-level in the full image canvas, which lets predictions for nearby regions properly compete for space. Finally, we compare to interactive single object segmentation on the COCO panoptic dataset. We demonstrate that our interactive full image segmentation approach leads to a 5% IoU gain, reaching 90% IoU at a budget of four extreme clicks and four corrective scribbles per region., Comment: Accepted to CVPR 2019

We propose the first practical learned lossless image compression system, L3C, and show that it outperforms the popular engineered codecs, PNG, WebP and JPEG 2000. At the core of our method is a fully parallelizable hierarchical probabilistic model for adaptive entropy coding which is optimized end-to-end for the compression task. In contrast to recent autoregressive discrete probabilistic models such as PixelCNN, our method i) models the image distribution jointly with learned auxiliary representations instead of exclusively modeling the image distribution in RGB space, and ii) only requires three forward-passes to predict all pixel probabilities instead of one for each pixel. As a result, L3C obtains over two orders of magnitude speedups when sampling compared to the fastest PixelCNN variant (Multiscale-PixelCNN). Furthermore, we find that learning the auxiliary representation is crucial and outperforms predefined auxiliary representations such as an RGB pyramid significantly., Comment: Updated preprocessing and Table 1, see A.1 in supplementary. Code and models: https://github.com/fab-jul/L3C-PyTorch

This paper reviews the first challenge on efficient perceptual image enhancement with the focus on deploying deep learning models on smartphones. The challenge consisted of two tracks. In the first one, participants were solving the classical image super-resolution problem with a bicubic downscaling factor of 4. The second track was aimed at real-world photo enhancement, and the goal was to map low-quality photos from the iPhone 3GS device to the same photos captured with a DSLR camera. The target metric used in this challenge combined the runtime, PSNR scores and solutions' perceptual results measured in the user study. To ensure the efficiency of the submitted models, we additionally measured their runtime and memory requirements on Android smartphones. The proposed solutions significantly improved baseline results defining the state-of-the-art for image enhancement on smartphones.

We propose and study the problem of distribution-preserving lossy compression. Motivated by recent advances in extreme image compression which allow to maintain artifact-free reconstructions even at very low bitrates, we propose to optimize the rate-distortion tradeoff under the constraint that the reconstructed samples follow the distribution of the training data. The resulting compression system recovers both ends of the spectrum: On one hand, at zero bitrate it learns a generative model of the data, and at high enough bitrates it achieves perfect reconstruction. Furthermore, for intermediate bitrates it smoothly interpolates between learning a generative model of the training data and perfectly reconstructing the training samples. We study several methods to approximately solve the proposed optimization problem, including a novel combination of Wasserstein GAN and Wasserstein Autoencoder, and present an extensive theoretical and empirical characterization of the proposed compression systems., Comment: NIPS 2018. Code: https://github.com/mitscha/dplc . Changes w.r.t. v1: Some clarifications in the text and additional numerical results

We present a learned image compression system based on GANs, operating at extremely low bitrates. Our proposed framework combines an encoder, decoder/generator and a multi-scale discriminator, which we train jointly for a generative learned compression objective. The model synthesizes details it cannot afford to store, obtaining visually pleasing results at bitrates where previous methods fail and show strong artifacts. Furthermore, if a semantic label map of the original image is available, our method can fully synthesize unimportant regions in the decoded image such as streets and trees from the label map, proportionally reducing the storage cost. A user study confirms that for low bitrates, our approach is preferred to state-of-the-art methods, even when they use more than double the bits., Comment: E. Agustsson, M. Tschannen, and F. Mentzer contributed equally to this work. ICCV 2019 camera ready version

Motivated by recent work on deep neural network (DNN)-based image compression methods showing potential improvements in image quality, savings in storage, and bandwidth reduction, we propose to perform image understanding tasks such as classification and segmentation directly on the compressed representations produced by these compression methods. Since the encoders and decoders in DNN-based compression methods are neural networks with feature-maps as internal representations of the images, we directly integrate these with architectures for image understanding. This bypasses decoding of the compressed representation into RGB space and reduces computational cost. Our study shows that accuracies comparable to networks that operate on compressed RGB images can be achieved while reducing the computational complexity up to $2\times$. Furthermore, we show that synergies are obtained by jointly training compression networks with classification networks on the compressed representations, improving image quality, classification accuracy, and segmentation performance. We find that inference from compressed representations is particularly advantageous compared to inference from compressed RGB images for aggressive compression rates., Comment: ICLR 2018 conference paper

Deep Neural Networks trained as image auto-encoders have recently emerged as a promising direction for advancing the state-of-the-art in image compression. The key challenge in learning such networks is twofold: To deal with quantization, and to control the trade-off between reconstruction error (distortion) and entropy (rate) of the latent image representation. In this paper, we focus on the latter challenge and propose a new technique to navigate the rate-distortion trade-off for an image compression auto-encoder. The main idea is to directly model the entropy of the latent representation by using a context model: A 3D-CNN which learns a conditional probability model of the latent distribution of the auto-encoder. During training, the auto-encoder makes use of the context model to estimate the entropy of its representation, and the context model is concurrently updated to learn the dependencies between the symbols in the latent representation. Our experiments show that this approach, when measured in MS-SSIM, yields a state-of-the-art image compression system based on a simple convolutional auto-encoder., Comment: CVPR 2018. Code available at https://github.com/fab-jul/imgcomp-cvpr . The first two authors contributed equally. Minor revision: fixed Fig. 2, added page numbers

This year alone has seen unprecedented leaps in the area of learning-based image translation, namely CycleGAN, by Zhu et al. But experiments so far have been tailored to merely two domains at a time, and scaling them to more would require an quadratic number of models to be trained. And with two-domain models taking days to train on current hardware, the number of domains quickly becomes limited by the time and resources required to process them. In this paper, we propose a multi-component image translation model and training scheme which scales linearly - both in resource consumption and time required - with the number of domains. We demonstrate its capabilities on a dataset of paintings by 14 different artists and on images of the four different seasons in the Alps. Note that 14 data groups would need (14 choose 2) = 91 different CycleGAN models: a total of 182 generator/discriminator pairs; whereas our model requires only 14 generator/discriminator pairs., Comment: Source code provided here: https://github.com/AAnoosheh/ComboGAN

Designing a logo for a new brand is a lengthy and tedious back-and-forth process between a designer and a client. In this paper we explore to what extent machine learning can solve the creative task of the designer. For this, we build a dataset -- LLD -- of 600k+ logos crawled from the world wide web. Training Generative Adversarial Networks (GANs) for logo synthesis on such multi-modal data is not straightforward and results in mode collapse for some state-of-the-art methods. We propose the use of synthetic labels obtained through clustering to disentangle and stabilize GAN training. We are able to generate a high diversity of plausible logos and we demonstrate latent space exploration techniques to ease the logo design task in an interactive manner. Moreover, we validate the proposed clustered GAN training on CIFAR 10, achieving state-of-the-art Inception scores when using synthetic labels obtained via clustering the features of an ImageNet classifier. GANs can cope with multi-modal data by means of synthetic labels achieved through clustering, and our results show the creative potential of such techniques for logo synthesis and manipulation. Our dataset and models will be made publicly available at https://data.vision.ee.ethz.ch/cvl/lld/.

Generative models such as Variational Auto Encoders (VAEs) and Generative Adversarial Networks (GANs) are typically trained for a fixed prior distribution in the latent space, such as uniform or Gaussian. After a trained model is obtained, one can sample the Generator in various forms for exploration and understanding, such as interpolating between two samples, sampling in the vicinity of a sample or exploring differences between a pair of samples applied to a third sample. In this paper, we show that the latent space operations used in the literature so far induce a distribution mismatch between the resulting outputs and the prior distribution the model was trained on. To address this, we propose to use distribution matching transport maps to ensure that such latent space operations preserve the prior distribution, while minimally modifying the original operation. Our experimental results validate that the proposed operations give higher quality samples compared to the original operations.

In this paper, we present a study on learning visual recognition models from large scale noisy web data. We build a new database called WebVision, which contains more than $2.4$ million web images crawled from the Internet by using queries generated from the 1,000 semantic concepts of the benchmark ILSVRC 2012 dataset. Meta information along with those web images (e.g., title, description, tags, etc.) are also crawled. A validation set and test set containing human annotated images are also provided to facilitate algorithmic development. Based on our new database, we obtain a few interesting observations: 1) the noisy web images are sufficient for training a good deep CNN model for visual recognition; 2) the model learnt from our WebVision database exhibits comparable or even better generalization ability than the one trained from the ILSVRC 2012 dataset when being transferred to new datasets and tasks; 3) a domain adaptation issue (a.k.a., dataset bias) is observed, which means the dataset can be used as the largest benchmark dataset for visual domain adaptation. Our new WebVision database and relevant studies in this work would benefit the advance of learning state-of-the-art visual models with minimum supervision based on web data.

We present the 2017 WebVision Challenge, a public image recognition challenge designed for deep learning based on web images without instance-level human annotation. Following the spirit of previous vision challenges, such as ILSVRC, Places2 and PASCAL VOC, which have played critical roles in the development of computer vision by contributing to the community with large scale annotated data for model designing and standardized benchmarking, we contribute with this challenge a large scale web images dataset, and a public competition with a workshop co-located with CVPR 2017. The WebVision dataset contains more than $2.4$ million web images crawled from the Internet by using queries generated from the $1,000$ semantic concepts of the benchmark ILSVRC 2012 dataset. Meta information is also included. A validation set and test set containing human annotated images are also provided to facilitate algorithmic development. The 2017 WebVision challenge consists of two tracks, the image classification task on WebVision test set, and the transfer learning task on PASCAL VOC 2012 dataset. In this paper, we describe the details of data collection and annotation, highlight the characteristics of the dataset, and introduce the evaluation metrics., Comment: project page: http://www.vision.ee.ethz.ch/webvision/

We present a new approach to learn compressible representations in deep architectures with an end-to-end training strategy. Our method is based on a soft (continuous) relaxation of quantization and entropy, which we anneal to their discrete counterparts throughout training. We showcase this method for two challenging applications: Image compression and neural network compression. While these tasks have typically been approached with different methods, our soft-to-hard quantization approach gives results competitive with the state-of-the-art for both.

We propose k^2-means, a new clustering method which efficiently copes with large numbers of clusters and achieves low energy solutions. k^2-means builds upon the standard k-means (Lloyd's algorithm) and combines a new strategy to accelerate the convergence with a new low time complexity divisive initialization. The accelerated convergence is achieved through only looking at k_n nearest clusters and using triangle inequality bounds in the assignment step while the divisive initialization employs an optimal 2-clustering along a direction. The worst-case time complexity per iteration of our k^2-means is O(nk_nd+k^2d), where d is the dimension of the n data points and k is the number of clusters and usually n << k << k_n. Compared to k-means' O(nkd) complexity, our k^2-means complexity is significantly lower, at the expense of slightly increasing the memory complexity by O(nk_n+k^2). In our extensive experiments k^2-means is order(s) of magnitude faster than standard methods in computing accurate clusterings on several standard datasets and settings with hundreds of clusters and high dimensional data. Moreover, the proposed divisive initialization generally leads to clustering energies comparable to those achieved with the standard k-means++ initialization, while being significantly faster.

We propose an information-theoretic framework for analog signal separation. Specifically, we consider the problem of recovering two analog signals, modeled as general random vectors, from the noiseless sum of linear measurements of the signals. Our framework is inspired by the groundbreaking work of Wu and Verd\'u (2010) on analog compression and encompasses, inter alia, inpainting, declipping, super-resolution, the recovery of signals corrupted by impulse noise, and the separation of (e.g., audio or video) signals into two distinct components. The main results we report are general achievability bounds for the compression rate, i.e., the number of measurements relative to the dimension of the ambient space the signals live in, under either measurability or H\"older continuity imposed on the separator. Furthermore, we find a matching converse for sources of mixed discrete-continuous distribution. For measurable separators our proofs are based on a new probabilistic uncertainty relation which shows that the intersection of generic subspaces with general sets of sufficiently small Minkowski dimension is empty. H\"older continuous separators are dealt with by introducing the concept of regularized probabilistic uncertainty relations. The probabilistic uncertainty relations we develop are inspired by embedding results in dynamical systems theory due to Sauer et al. (1991) and---conceptually---parallel classical Donoho-Stark and Elad-Bruckstein uncertainty principles at the heart of compressed sensing theory. Operationally, the new uncertainty relations take the theory of sparse signal separation beyond traditional sparsity---as measured in terms of the number of non-zero entries---to the more general notion of low description complexity as quantified by Minkowski dimension. Finally, our approach also allows to significantly strengthen key results in Wu and Verd\'u (2010)., Comment: to appear in IEEE Trans. on Inf. Theory

Subspace clustering refers to the problem of clustering high-dimensional data points into a union of low-dimensional linear subspaces, where the number of subspaces, their dimensions and orientations are all unknown. In this paper, we propose a variation of the recently introduced thresholding-based subspace clustering (TSC) algorithm, which applies spectral clustering to an adjacency matrix constructed from the nearest neighbors of each data point with respect to the spherical distance measure. The new element resides in an individual and data-driven choice of the number of nearest neighbors. Previous performance results for TSC, as well as for other subspace clustering algorithms based on spectral clustering, come in terms of an intermediate performance measure, which does not address the clustering error directly. Our main analytical contribution is a performance analysis of the modified TSC algorithm (as well as the original TSC algorithm) in terms of the clustering error directly., Comment: ICASSP 2014