We propose a discrete latent distribution for Generative Adversarial Networks (GANs). Instead of drawing latent vectors from a continuous prior, we sample from a finite set of learnable latents. However, a direct parametrization of such a distribution leads to an intractable linear increase in memory in order to ensure sufficient sample diversity. We address this key issue by taking inspiration from the encoding of information in biological organisms. Instead of learning a separate latent vector for each sample, we split the latent space into a set of genes. For each gene, we train a small bank of gene variants. Thus, by independently sampling a variant for each gene and combining them into the final latent vector, our approach can represent a vast number of unique latent samples from a compact set of learnable parameters. Interestingly, our gene-inspired latent encoding allows for new and intuitive approaches to latent-space exploration, enabling conditional sampling from our unconditionally trained model. Moreover, our approach preserves state-of-the-art photo-realism while achieving better disentanglement than the widely-used StyleMapping network.

We propose a novel edge guided generative adversarial network with contrastive learning (ECGAN) for the challenging semantic image synthesis task. Although considerable improvement has been achieved, the quality of synthesized images is far from satisfactory due to three largely unresolved challenges. 1) The semantic labels do not provide detailed structural information, making it difficult to synthesize local details and structures. However, they ignore "global" semantic information of multiple input semantic layouts, i.e., semantic cross-relations between pixels across different input layouts. To tackle 1), we propose to use edge as an intermediate representation which is further adopted to guide image generation via a proposed attention guided edge transfer module. Edge information is produced by a convolutional generator and introduces detailed structure information. To tackle 2), we design an effective module to selectively highlight class-dependent feature maps according to the original semantic layout to preserve the semantic information. To tackle 3), inspired by current methods in contrastive learning, we propose a novel contrastive learning method, which aims to enforce pixel embeddings belonging to the same semantic class to generate more similar image content than those from different classes. Doing so can capture more semantic relations by explicitly exploring the structures of labeled pixels from multiple input semantic layouts. Experiments on three challenging datasets show that our ECGAN achieves significantly better results than state-of-the-art methods.

We present a novel graph Transformer generative adversarial network (GTGAN) to learn effective graph node relations in an end-to-end fashion for the challenging graph-constrained house generation task. The proposed graph-Transformer-based generator includes a novel graph Transformer encoder that combines graph convolutions and self-attentions in a Transformer to model both local and global interactions across connected and non-connected graph nodes. Specifically, the proposed connected node attention (CNA) and non-connected node attention (NNA) aim to capture the global relations across connected nodes and non-connected nodes in the input graph, respectively. The proposed graph modeling block (GMB) aims to exploit local vertex interactions based on a house layout topology. Moreover, we propose a new node classification-based discriminator to preserve the high-level semantic and discriminative node features for different house components. Finally, we propose a novel graph-based cycle-consistency loss that aims at maintaining the relative spatial relationships between ground truth and predicted graphs. Experiments on two challenging graph-constrained house generation tasks (i.e., house layout and roof generation) with two public datasets demonstrate the effectiveness of GTGAN in terms of objective quantitative scores and subjective visual realism. New state-of-the-art results are established by large margins on both tasks.

We introduce VA-DepthNet, a simple, effective, and accurate deep neural network approach for the single-image depth prediction (SIDP) problem. The proposed approach advocates using classical first-order variational constraints for this problem. While state-of-the-art deep neural network methods for SIDP learn the scene depth from images in a supervised setting, they often overlook the invaluable invariances and priors in the rigid scene space, such as the regularity of the scene. The paper's main contribution is to reveal the benefit of classical and well-founded variational constraints in the neural network design for the SIDP task. It is shown that imposing first-order variational constraints in the scene space together with popular encoder-decoder-based network architecture design provides excellent results for the supervised SIDP task. The imposed first-order variational constraint makes the network aware of the depth gradient in the scene space, i.e., regularity. The paper demonstrates the usefulness of the proposed approach via extensive evaluation and ablation analysis over several benchmark datasets, such as KITTI, NYU Depth V2, and SUN RGB-D. The VA-DepthNet at test time shows considerable improvements in depth prediction accuracy compared to the prior art and is accurate also at high-frequency regions in the scene space. Over the last decade, neural networks have introduced a new prospect for the 3D computer vision field. It has led to significant progress on many long-standing problems in this field, such as multiview stereo (Huang et al., 2018; Kaya et al., 2022), visual simultaneous localization and mapping (Teed & Deng, 2021), novel view synthesis (Mildenhall et al., 2021), etc. Among several 3D vision problems, one of the challenging, if not impossible, to solve is the single-image depth prediction (SIDP) problem. SIDP is indeed ill-posed--in a strict geometric sense, presenting an extraordinary challenge to solve this inverse problem reliably. Moreover, since we do not have access to multi-view images, it is hard to constrain this problem via well-known geometric constraints (Longuet-Higgins, 1981; Nistér, 2004; Furukawa & Ponce, 2009; Kumar et al., 2019; 2017). Accordingly, the SIDP problem generally boils down to an ambitious fitting problem, to which deep learning provides a suitable way to predict an acceptable solution to this problem (Yuan et al., 2022; Yin et al., 2019). Impressive earlier methods use Markov Random Fields (MRF) to model monocular cues and the relation between several over-segmented image parts (Saxena et al., 2007; 2008). Popular recent methods for SIDP are mostly supervised.

End-to-end Automatic Speech Recognition (ASR) systems based on neural networks have seen large improvements in recent years. The availability of large scale hand-labeled datasets and sufficient computing resources made it possible to train powerful deep neural networks, reaching very low Word Error Rate (WER) on academic benchmarks. However, despite impressive performance on clean audio samples, a drop of performance is often observed on noisy speech. In this work, we propose to improve the noise robustness of the recently proposed Efficient Conformer Connectionist Temporal Classification (CTC)-based architecture by processing both audio and visual modalities. We improve previous lip reading methods using an Efficient Conformer back-end on top of a ResNet-18 visual front-end and by adding intermediate CTC losses between blocks. We condition intermediate block features on early predictions using Inter CTC residual modules to relax the conditional independence assumption of CTC-based models. We also replace the Efficient Conformer grouped attention by a more efficient and simpler attention mechanism that we call patch attention. We experiment with publicly available Lip Reading Sentences 2 (LRS2) and Lip Reading Sentences 3 (LRS3) datasets. Our experiments show that using audio and visual modalities allows to better recognize speech in the presence of environmental noise and significantly accelerate training, reaching lower WER with 4 times less training steps. Our Audio-Visual Efficient Conformer (AVEC) model achieves state-of-the-art performance, reaching WER of 2.3% and 1.8% on LRS2 and LRS3 test sets. Code and pretrained models are available at https://github.com/burchim/AVEC.

While neural networks-based photo processing solutions can provide a better image quality compared to the traditional ISP systems, their application to mobile devices is still very limited due to their very high computational complexity. In this paper, we present a novel MicroISP model designed specifically for edge devices, taking into account their computational and memory limitations. The proposed solution is capable of processing up to 32MP photos on recent smartphones using the standard mobile ML libraries and requiring less than 1 second to perform the inference, while for FullHD images it achieves real-time performance. The architecture of the model is flexible, allowing to adjust its complexity to devices of different computational power. To evaluate the performance of the model, we collected a novel Fujifilm UltraISP dataset consisting of thousands of paired photos captured with a normal mobile camera sensor and a professional 102MP medium-format FujiFilm GFX100 camera. The experiments demonstrated that, despite its compact size, the MicroISP model is able to provide comparable or better visual results than the traditional mobile ISP systems, while outperforming the previously proposed efficient deep learning based solutions. Finally, this model is also compatible with the latest mobile AI accelerators, achieving good runtime and low power consumption on smartphone NPUs and APUs. The code, dataset and pre-trained models are available on the project website: https://people.ee.ethz.ch/~ihnatova/microisp.html

Shadow removal is an important computer vision task aiming at the detection and successful removal of the shadow produced by an occluded light source and a photo-realistic restoration of the image contents. Decades of re-search produced a multitude of hand-crafted restoration techniques and, more recently, learned solutions from shad-owed and shadow-free training image pairs. In this work,we propose an unsupervised single image shadow removal solution via self-supervised learning by using a conditioned mask. In contrast to existing literature, we do not require paired shadowed and shadow-free images, instead we rely on self-supervision and jointly learn deep models to remove and add shadows to images. We validate our approach on the recently introduced ISTD and USR datasets. We largely improve quantitatively and qualitatively over the compared methods and set a new state-of-the-art performance in single image shadow removal.

Energy-based models (EBMs) have become increasingly popular within computer vision in recent years. While they are commonly employed for generative image modeling, recent work has applied EBMs also for regression tasks, achieving state-of-the-art performance on object detection and visual tracking. Training EBMs is however known to be challenging. While a variety of different techniques have been explored for generative modeling, the application of EBMs to regression is not a well-studied problem. How EBMs should be trained for best possible regression performance is thus currently unclear. We therefore accept the task of providing the first detailed study of this problem. To that end, we propose a simple yet highly effective extension of noise contrastive estimation, and carefully compare its performance to six popular methods from literature on the tasks of 1D regression and object detection. The results of this comparison suggest that our training method should be considered the go-to approach. We also apply our method to the visual tracking task, achieving state-of-the-art performance on five datasets. Notably, our tracker achieves 63.7% AUC on LaSOT and 78.7% Success on TrackingNet. Code is available at https://github.com/fregu856/ebms_regression.

This paper introduces a divide-and-conquer inspired adversarial learning (DA-CAL) approach for photo enhancement. The key idea is to decompose the photo enhancement process into hierarchically multiple sub-problems, which can be better conquered from bottom to up. On the top level, we propose a perception-based division to learn additive and multiplicative components, required to translate a low-quality image or video into its high-quality counterpart. On the intermediate level, we use a frequency-based division with generative adversarial network (GAN) to weakly supervise the photo enhancement process. On the lower level, we design a dimension-based division that enables the GAN model to better approximates the distribution distance on multiple independent one-dimensional data to train the GAN model. While considering all three hierarchies, we develop multiscale and recurrent training approaches to optimize the image and video enhancement process in a weakly-supervised manner. Both quantitative and qualitative results clearly demonstrate that the proposed DACAL achieves the state-of- the-art performance for high-resolution image and video enhancement. Despite many mobile camera technological advances we have today, our captured images often still come with limited dynamic range, undesirable color rendition, and unsatisfactory texture sharpness. Among many possible causes, low-light environments and under/overexposed regions usually introduce severe lack of texture details and low-dynamic range coverage, respectively. Another critical issue is the amplification (during the enhancement process) of noise in the dark and/or texture-less regions, where the enhancement may not even be necessary.

Over the last years, the computational power of mobile devices such as smartphones and tablets has grown dramatically, reaching the level of desktop computers available not long ago. While standard smartphone apps are no longer a problem for them, there is still a group of tasks that can easily challenge even high-end devices, namely running artificial intelligence algorithms. In this paper, we present a study of the current state of deep learning in the Android ecosystem and describe available frameworks, programming models and the limitations of running AI on smartphones. We give an overview of the hardware acceleration resources available on four main mobile chipset platforms: Qualcomm, HiSilicon, MediaTek and Samsung. Additionally, we present the real-world performance results of different mobile SoCs collected with AI Benchmark that are covering all main existing hardware configurations.