Under-display camera (UDC) provides an elegant solution for full-screen smartphones. However, UDC captured images suffer from severe degradation since sensors lie under the display. Although this issue can be tackled by image restoration networks, these networks require large-scale image pairs for training. To this end, we propose a modular network dubbed MPGNet trained using the generative adversarial network (GAN) framework for simulating UDC imaging. Specifically, we note that the UDC imaging degradation process contains brightness attenuation, blurring, and noise corruption. Thus we model each degradation with a characteristic-related modular network, and all modular networks are cascaded to form the generator. Together with a pixel-wise discriminator and supervised loss, we can train the generator to simulate the UDC imaging degradation process. Furthermore, we present a Transformer-style network named DWFormer for UDC image restoration. For practical purposes, we use depth-wise convolution instead of the multi-head self-attention to aggregate local spatial information. Moreover, we propose a novel channel attention module to aggregate global information, which is critical for brightness recovery. We conduct evaluations on the UDC benchmark, and our method surpasses the previous state-of-the-art models by 1.23 dB on the P-OLED track and 0.71 dB on the T-OLED track, respectively.
In this work, we further develop the conformer-based metric generative adversarial network (CMGAN) model for speech enhancement (SE) in the time-frequency (TF) domain. This paper builds on our previous work but takes a more in-depth look by conducting extensive ablation studies on model inputs and architectural design choices. We rigorously tested the generalization ability of the model to unseen noise types and distortions. We have fortified our claims through DNS-MOS measurements and listening tests. Rather than focusing exclusively on the speech denoising task, we extend this work to address the dereverberation and super-resolution tasks. This necessitated exploring various architectural changes, specifically metric discriminator scores and masking techniques. It is essential to highlight that this is among the earliest works that attempted complex TF-domain super-resolution. Our findings show that CMGAN outperforms existing state-of-the-art methods in the three major speech enhancement tasks: denoising, dereverberation, and super-resolution. For example, in the denoising task using the Voice Bank+DEMAND dataset, CMGAN notably exceeded the performance of prior models, attaining a PESQ score of 3.41 and an SSNR of 11.10 dB. Audio samples and CMGAN implementations are available online.
Singing voice synthesis (SVS) is the computer production of a human-like singing voice from given musical scores. To accomplish end-to-end SVS effectively and efficiently, this work adopts the acoustic model-neural vocoder architecture established for high-quality speech and singing voice synthesis. Specifically, this work aims to pursue a higher level of expressiveness in synthesized voices by combining the diffusion denoising probabilistic model (DDPM) and \emph{Wasserstein} generative adversarial network (WGAN) to construct the backbone of the acoustic model. On top of the proposed acoustic model, a HiFi-GAN neural vocoder is adopted with integrated fine-tuning to ensure optimal synthesis quality for the resulting end-to-end SVS system. This end-to-end system was evaluated with the multi-singer Mpop600 Mandarin singing voice dataset. In the experiments, the proposed system exhibits improvements over previous landmark counterparts in terms of musical expressiveness and high-frequency acoustic details. Moreover, the adversarial acoustic model converged stably without the need to enforce reconstruction objectives, indicating the convergence stability of the proposed DDPM and WGAN combined architecture over alternative GAN-based SVS systems.
Although recent gaze estimation methods lay great emphasis on attentively extracting gaze-relevant features from facial or eye images, how to define features that include gaze-relevant components has been ambiguous. This obscurity makes the model learn not only gaze-relevant features but also irrelevant ones. In particular, it is fatal for the cross-dataset performance. To overcome this challenging issue, we propose a gaze-aware analytic manipulation method, based on a data-driven approach with generative adversarial network inversion's disentanglement characteristics, to selectively utilize gaze-relevant features in a latent code. Furthermore, by utilizing GAN-based encoder-generator process, we shift the input image from the target domain to the source domain image, which a gaze estimator is sufficiently aware. In addition, we propose gaze distortion loss in the encoder that prevents the distortion of gaze information. The experimental results demonstrate that our method achieves state-of-the-art gaze estimation accuracy in a cross-domain gaze estimation tasks. This code is available at https://github.com/leeisack/LatentGaze/.
The goal of this work is to help mitigate the already existing gender wage gap by supplying unbiased job recommendations based on resumes from job seekers. We employ a generative adversarial network to remove gender bias from word2vec representations of 12M job vacancy texts and 900k resumes. Our results show that representations created from recruitment texts contain algorithmic bias and that this bias results in real-world consequences for recommendation systems. Without controlling for bias, women are recommended jobs with significantly lower salary in our data. With adversarially fair representations, this wage gap disappears, meaning that our debiased job recommendations reduce wage discrimination. We conclude that adversarial debiasing of word representations can increase real-world fairness of systems and thus may be part of the solution for creating fairness-aware recommendation systems.
To recognize the masked face, one of the possible solutions could be to restore the occluded part of the face first and then apply the face recognition method. Inspired by the recent image inpainting methods, we propose an end-to-end hybrid masked face recognition system, namely HiMFR, consisting of three significant parts: masked face detector, face inpainting, and face recognition. The masked face detector module applies a pretrained Vision Transformer (ViT\_b32) to detect whether faces are covered with masked or not. The inpainting module uses a fine-tune image inpainting model based on a Generative Adversarial Network (GAN) to restore faces. Finally, the hybrid face recognition module based on ViT with an EfficientNetB3 backbone recognizes the faces. We have implemented and evaluated our proposed method on four different publicly available datasets: CelebA, SSDMNV2, MAFA, {Pubfig83} with our locally collected small dataset, namely Face5. Comprehensive experimental results show the efficacy of the proposed HiMFR method with competitive performance. Code is available at https://github.com/mdhosen/HiMFR
Parallel imaging accelerates MRI data acquisition by acquiring additional sensitivity information with an array of receiver coils, resulting in fewer phase encoding steps. Because of fewer data requirements than parallel imaging, compressed sensing magnetic resonance imaging (CS-MRI) has gained popularity in the field of medical imaging. Parallel imaging and compressed sensing (CS) both reduce the amount of data captured in the k-space, which speeds up traditional MRI acquisition. As acquisition time is inversely proportional to sample count, forming an image from reduced k-space samples results in faster acquisition but with aliasing artifacts. For de-aliasing the reconstructed image, this paper proposes a novel Generative Adversarial Network (GAN) called RECGAN-GR that is supervised with multi-modal losses. In comparison to existing GAN networks, our proposed method introduces a novel generator network, RemU-Net, which is integrated with dual-domain loss functions such as weighted magnitude and phase loss functions, as well as parallel imaging-based loss, GRAPPA consistency loss. As refinement learning, a k-space correction block is proposed to make the GAN network self-resistant to generating unnecessary data, which speeds up the reconstruction process. Comprehensive results show that the proposed RECGAN-GR not only improves the PSNR by 4 dB over GAN-based methods but also by 2 dB over conventional state-of-the-art CNN methods available in the literature for single-coil data. The proposed work significantly improves image quality for low-retained data, resulting in five to ten times faster acquisition.
In this paper, we present the first conditional generative adversarial network (cGAN) architecture that is specifically designed to fuse synthetic aperture radar (SAR) and optical multi-spectral (MS) image data to generate cloud- and haze-free MS optical data from a cloud-corrupted MS input and an auxiliary SAR image. Experiments on Sentinel-2 MS and Sentinel-l SAR data confirm that our extended SAR-Opt-cGAN model utilizes the auxiliary SAR information to better reconstruct MS images than an equivalent model which uses the same architecture but only single-sensor MS data as input.
Deep reinforcement learning has demonstrated superhuman performance in complex decision-making tasks, but it struggles with generalization and knowledge reuse - key aspects of true intelligence. This article introduces a novel approach that modifies Cycle Generative Adversarial Networks specifically for reinforcement learning, enabling effective one-to-one knowledge transfer between two tasks. Our method enhances the loss function with two new components: model loss, which captures dynamic relationships between source and target tasks, and Q-loss, which identifies states significantly influencing the target decision policy. Tested on the 2-D Atari game Pong, our method achieved 100% knowledge transfer in identical tasks and either 100% knowledge transfer or a 30% reduction in training time for a rotated task, depending on the network architecture. In contrast, using standard Generative Adversarial Networks or Cycle Generative Adversarial Networks led to worse performance than training from scratch in the majority of cases. The results demonstrate that the proposed method ensured enhanced knowledge generalization in deep reinforcement learning.
Deep learning-based techniques have been introduced into the field of trajectory optimization in recent years. Deep Neural Networks (DNNs) are trained and used as the surrogates of conventional optimization process. They can provide low thrust (LT) transfer cost estimation and enable more complex preliminary mission designs. However, it is a challenge to efficiently obtain the required amount of trajectory data for training. A Generative Adversarial Network (GAN) is adapted to generate the feasible LT trajectory data efficiently. The GAN consists of a generator and a discriminator, both of which are deep networks. The generator generates fake LT transfer features using random noise as input, while the discriminator distinguishes the generator's fake LT transfer features from real LT transfer features. The GAN is trained until the generator generates fake LT transfers that the discriminator cannot identify. This indicates the generator generates low thrust transfer features that have the same distribution as the real transfer features. The generated low thrust transfer data have a high convergence rate, and they can be used to efficiently produce training data for deep learning models. The proposed approach is validated by generating feasible LT transfers in a Near-Earth Asteroid (NEA) mission scenario. The convergence rate of GAN-generated samples is 84.3%.
Clinical screening with low-quality fundus images is challenging and significantly leads to misdiagnosis. This paper addresses the issue of improving the retinal image quality and vessel segmentation through retinal image restoration. More specifically, a cycle-consistent generative adversarial network (CycleGAN) with a convolution block attention module (CBAM) is used for retinal image restoration. A modified UNet is used for retinal vessel segmentation for the restored retinal images (CBAM-UNet). The proposed model consists of two generators and two discriminators. Generators translate images from one domain to another, i.e., from low to high quality and vice versa. Discriminators classify generated and original images. The retinal vessel segmentation model uses downsampling, bottlenecking, and upsampling layers to generate segmented images. The CBAM has been used to enhance the feature extraction of these models. The proposed method does not require paired image datasets, which are challenging to produce. Instead, it uses unpaired data that consists of low- and high-quality fundus images retrieved from publicly available datasets. The restoration performance of the proposed method was evaluated using full-reference evaluation metrics, e.g., peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM). The retinal vessel segmentation performance was compared with the ground-truth fundus images. The proposed method can significantly reduce the degradation effects caused by out-of-focus blurring, color distortion, low, high, and uneven illumination. Experimental results show the effectiveness of the proposed method for retinal image restoration and vessel segmentation.
Many activity classifications segments data into fixed window size for feature extraction and classification. However, animal behaviors have various durations that do not match the predetermined window size. The dense labeling and dense prediction methods address this limitation by predicting labels for every point. Thus, by tracing the starting and ending points, we could know the time location and duration of all occurring activities. Still, the dense prediction could be noisy with misalignments problems. We modified the U-Net and Conditional Generative Adversarial Network (cGAN) with customized loss functions as a training strategy to reduce fragmentation and other misalignments. In cGAN, the discriminator and generator trained against each other like an adversarial competition. The generator produces dense predictions. The discriminator works as a high-level consistency check, in our case, pushing the generator to predict activities with reasonable duration. The model trained with cGAN shows better or comparable performance in the cow, pig, and UCI HAPT dataset. The cGAN-trained modified U-Net improved from 92.17% to 94.66% for the UCI HAPT dataset and from 90.85% to 93.18% for pig data compared to previous dense prediction work.
Due to the rapid growth of Electrical Capacitance Tomography (ECT) applications in several industrial fields, there is a crucial need for developing high quality, yet fast, methodologies of image reconstruction from raw capacitance measurements. Deep learning, as an effective non-linear mapping tool for complicated functions, has been going viral in many fields including electrical tomography. In this paper, we propose a Conditional Generative Adversarial Network (CGAN) model for reconstructing ECT images from capacitance measurements. The initial image of the CGAN model is constructed from the capacitance measurement. To our knowledge, this is the first time to represent the capacitance measurements in an image form. We have created a new massive ECT dataset of 320K synthetic image measurements pairs for training, and testing the proposed model. The feasibility and generalization ability of the proposed CGAN-ECT model are evaluated using testing dataset, contaminated data and flow patterns that are not exposed to the model during the training phase. The evaluation results prove that the proposed CGAN-ECT model can efficiently create more accurate ECT images than traditional and other deep learning-based image reconstruction algorithms. CGAN-ECT achieved an average image correlation coefficient of more than 99.3% and an average relative image error about 0.07.
Single-Image Super-Resolution can support robotic tasks in environments where a reliable visual stream is required to monitor the mission, handle teleoperation or study relevant visual details. In this work, we propose an efficient Generative Adversarial Network model for real-time Super-Resolution, called EdgeSRGAN (code available at https://github.com/PIC4SeR/EdgeSRGAN). We adopt a tailored architecture of the original SRGAN and model quantization to boost the execution on CPU and Edge TPU devices, achieving up to 200 fps inference. We further optimize our model by distilling its knowledge to a smaller version of the network and obtain remarkable improvements compared to the standard training approach. Our experiments show that our fast and lightweight model preserves considerably satisfying image quality compared to heavier state-of-the-art models. Finally, we conduct experiments on image transmission with bandwidth degradation to highlight the advantages of the proposed system for mobile robotic applications.
The adaptation of a Generative Adversarial Network (GAN) aims to transfer a pre-trained GAN to a target domain with limited training data. In this paper, we focus on the one-shot case, which is more challenging and rarely explored in previous works. We consider that the adaptation from a source domain to a target domain can be decoupled into two parts: the transfer of global style like texture and color, and the emergence of new entities that do not belong to the source domain. While previous works mainly focus on style transfer, we propose a novel and concise framework to address the \textit{generalized one-shot adaptation} task for both style and entity transfer, in which a reference image and its binary entity mask are provided. Our core idea is to constrain the gap between the internal distributions of the reference and syntheses by sliced Wasserstein distance. To better achieve it, style fixation is used at first to roughly obtain the exemplary style, and an auxiliary network is introduced to the generator to disentangle entity and style transfer. Besides, to realize cross-domain correspondence, we propose the variational Laplacian regularization to constrain the smoothness of the adapted generator. Both quantitative and qualitative experiments demonstrate the effectiveness of our method in various scenarios. Code is available at \url{https://github.com/zhangzc21/Generalized-One-shot-GAN-adaptation}.
A face morph is created by strategically combining two or more face images corresponding to multiple identities. The intention is for the morphed image to match with multiple identities. Current morph attack detection strategies can detect morphs but cannot recover the images or identities used in creating them. The task of deducing the individual face images from a morphed face image is known as \textit{de-morphing}. Existing work in de-morphing assume the availability of a reference image pertaining to one identity in order to recover the image of the accomplice - i.e., the other identity. In this work, we propose a novel de-morphing method that can recover images of both identities simultaneously from a single morphed face image without needing a reference image or prior information about the morphing process. We propose a generative adversarial network that achieves single image-based de-morphing with a surprisingly high degree of visual realism and biometric similarity with the original face images. We demonstrate the performance of our method on landmark-based morphs and generative model-based morphs with promising results.
Art is an artistic method of using digital technologies as a part of the generative or creative process. With the advent of digital currency and NFTs (Non-Fungible Token), the demand for digital art is growing aggressively. In this manuscript, we advocate the concept of using deep generative networks with adversarial training for a stable and variant art generation. The work mainly focuses on using the Deep Convolutional Generative Adversarial Network (DC-GAN) and explores the techniques to address the common pitfalls in GAN training. We compare various architectures and designs of DC-GANs to arrive at a recommendable design choice for a stable and realistic generation. The main focus of the work is to generate realistic images that do not exist in reality but are synthesised from random noise by the proposed model. We provide visual results of generated animal face images (some pieces of evidence showing a blend of species) along with recommendations for training, architecture and design choices. We also show how training image preprocessing plays a massive role in GAN training.
Distinguishing between computer-generated (CG) and natural photographic (PG) images is of great importance to verify the authenticity and originality of digital images. However, the recent cutting-edge generation methods enable high qualities of synthesis in CG images, which makes this challenging task even trickier. To address this issue, a joint learning strategy with deep texture and high-frequency features for CG image detection is proposed. We first formulate and deeply analyze the different acquisition processes of CG and PG images. Based on the finding that multiple different modules in image acquisition will lead to different sensitivity inconsistencies to the convolutional neural network (CNN)-based rendering in images, we propose a deep texture rendering module for texture difference enhancement and discriminative texture representation. Specifically, the semantic segmentation map is generated to guide the affine transformation operation, which is used to recover the texture in different regions of the input image. Then, the combination of the original image and the high-frequency components of the original and rendered images are fed into a multi-branch neural network equipped with attention mechanisms, which refines intermediate features and facilitates trace exploration in spatial and channel dimensions respectively. Extensive experiments on two public datasets and a newly constructed dataset with more realistic and diverse images show that the proposed approach outperforms existing methods in the field by a clear margin. Besides, results also demonstrate the detection robustness and generalization ability of the proposed approach to postprocessing operations and generative adversarial network (GAN) generated images.
Low resolution of positron emission tomography (PET) limits its diagnostic performance. Deep learning has been successfully applied to achieve super-resolution PET. However, commonly used supervised learning methods in this context require many pairs of low- and high-resolution (LR and HR) PET images. Although unsupervised learning utilizes unpaired images, the results are not as good as that obtained with supervised deep learning. In this paper, we propose a quasi-supervised learning method, which is a new type of weakly-supervised learning methods, to recover HR PET images from LR counterparts by leveraging similarity between unpaired LR and HR image patches. Specifically, LR image patches are taken from a patient as inputs, while the most similar HR patches from other patients are found as labels. The similarity between the matched HR and LR patches serves as a prior for network construction. Our proposed method can be implemented by designing a new network or modifying an existing network. As an example in this study, we have modified the cycle-consistent generative adversarial network (CycleGAN) for super-resolution PET. Our numerical and experimental results qualitatively and quantitatively show the merits of our method relative to the state-ofthe-art methods. The code is publicly available at https://github.com/PigYang-ops/CycleGAN-QSDL.
Automatic Speaker Verification (ASV) technology has become commonplace in virtual assistants. However, its performance suffers when there is a mismatch between the train and test domains. Mixed bandwidth training, i.e., pooling training data from both domains, is a preferred choice for developing a universal model that works for both narrowband and wideband domains. We propose complementing this technique by performing neural upsampling of narrowband signals, also known as bandwidth extension. Our main goal is to discover and analyze high-performing time-domain Generative Adversarial Network (GAN) based models to improve our downstream state-of-the-art ASV system. We choose GANs since they (1) are powerful for learning conditional distribution and (2) allow flexible plug-in usage as a pre-processor during the training of downstream task (ASV) with data augmentation. Prior works mainly focus on feature-domain bandwidth extension and limited experimental setups. We address these limitations by 1) using time-domain extension models, 2) reporting results on three real test sets, 2) extending training data, and 3) devising new test-time schemes. We compare supervised (conditional GAN) and unsupervised GANs (CycleGAN) and demonstrate average relative improvement in Equal Error Rate of 8.6% and 7.7%, respectively. For further analysis, we study changes in spectrogram visual quality, audio perceptual quality, t-SNE embeddings, and ASV score distributions. We show that our bandwidth extension leads to phenomena such as a shift of telephone (test) embeddings towards wideband (train) signals, a negative correlation of perceptual quality with downstream performance, and condition-independent score calibration.