Chao Wang

Chao Wang, Ana Serrano, Xingang Pan et al.

Most in-the-wild images are stored in Low Dynamic Range (LDR) form, serving as a partial observation of the High Dynamic Range (HDR) visual world. Despite limited dynamic range, these LDR images are often captured with different exposures, implicitly containing information about the underlying HDR image distribution. Inspired by this intuition, in this work we present, to the best of our knowledge, the first method for learning a generative model of HDR images from in-the-wild LDR image collections in a fully unsupervised manner. The key idea is to train a generative adversarial network (GAN) to generate HDR images which, when projected to LDR under various exposures, are indistinguishable from real LDR images. The projection from HDR to LDR is achieved via a camera model that captures the stochasticity in exposure and camera response function. Experiments show that our method GlowGAN can synthesize photorealistic HDR images in many challenging cases such as landscapes, lightning, or windows, where previous supervised generative models produce overexposed images. We further demonstrate the new application of unsupervised inverse tone mapping (ITM) enabled by GlowGAN. Our ITM method does not need HDR images or paired multi-exposure images for training, yet it reconstructs more plausible information for overexposed regions than state-of-the-art supervised learning models trained on such data.

Chao Wang

Multi-turn compositional image generation (M-CIG) is a challenging task that aims to iteratively manipulate a reference image given a modification text. While most of the existing methods for M-CIG are based on generative adversarial networks (GANs), recent advances in image generation have demonstrated the superiority of diffusion models over GANs. In this paper, we propose a diffusion-based method for M-CIG named conditional denoising diffusion with image compositional matching (CDD-ICM). We leverage CLIP as the backbone of image and text encoders, and incorporate a gated fusion mechanism, originally proposed for question answering, to compositionally fuse the reference image and the modification text at each turn of M-CIG. We introduce a conditioning scheme to generate the target image based on the fusion results. To prioritize the semantic quality of the generated target image, we learn an auxiliary image compositional match (ICM) objective, along with the conditional denoising diffusion (CDD) objective in a multi-task learning framework. Additionally, we also perform ICM guidance and classifier-free guidance to improve performance. Experimental results show that CDD-ICM achieves state-of-the-art results on two benchmark datasets for M-CIG, i.e., CoDraw and i-CLEVR.

Gustavo Aguilar, Viktor Rozgić, Weiran Wang et al.

Studies on emotion recognition (ER) show that combining lexical and acoustic information results in more robust and accurate models. The majority of the studies focus on settings where both modalities are available in training and evaluation. However, in practice, this is not always the case; getting ASR output may represent a bottleneck in a deployment pipeline due to computational complexity or privacy-related constraints. To address this challenge, we study the problem of efficiently combining acoustic and lexical modalities during training while still providing a deployable acoustic model that does not require lexical inputs. We first experiment with multimodal models and two attention mechanisms to assess the extent of the benefits that lexical information can provide. Then, we frame the task as a multi-view learning problem to induce semantic information from a multimodal model into our acoustic-only network using a contrastive loss function. Our multimodal model outperforms the previous state of the art on the USC-IEMOCAP dataset reported on lexical and acoustic information. Additionally, our multi-view-trained acoustic network significantly surpasses models that have been exclusively trained with acoustic features.

Ziqiang Zheng, Chao Wang, Zhibin Yu et al.

Semantic layouts based Image synthesizing, which has benefited from the success of Generative Adversarial Network (GAN), has drawn much attention in these days. How to enhance the synthesis image equality while keeping the stochasticity of the GAN is still a challenge. We propose a novel denoising framework to handle this problem. The overlapped objects generation is another challenging task when synthesizing images from a semantic layout to a realistic RGB photo. To overcome this deficiency, we include a one-hot semantic label map to force the generator paying more attention on the overlapped objects generation. Furthermore, we improve the loss function of the discriminator by considering perturb loss and cascade layer loss to guide the generation process. We applied our methods on the Cityscapes, Facades and NYU datasets and demonstrate the image generation ability of our model.

Ziqiang Zheng, Zhibin Yu, Haiyong Zheng et al.

Generative Adversarial Networks are proved to be efficient on various kinds of image generation tasks. However, it is still a challenge if we want to generate images precisely. Many researchers focus on how to generate images with one attribute. But image generation under multiple attributes is still a tough work. In this paper, we try to generate a variety of face images under multiple constraints using a pipeline process. The Pip-GAN (Pipeline Generative Adversarial Network) we present employs a pipeline network structure which can generate a complex facial image step by step using a neutral face image. We applied our method on two face image databases and demonstrate its ability to generate convincing novel images of unseen identities under multiple conditions previously.

Ziqiang Zheng, Wang Chao, Zhibin Yu et al.

Recently, image-to-image translation has been made much progress owing to the success of conditional Generative Adversarial Networks (cGANs). And some unpaired methods based on cycle consistency loss such as DualGAN, CycleGAN and DiscoGAN are really popular. However, it's still very challenging for translation tasks with the requirement of high-level visual information conversion, such as photo-to-caricature translation that requires satire, exaggeration, lifelikeness and artistry. We present an approach for learning to translate faces in the wild from the source photo domain to the target caricature domain with different styles, which can also be used for other high-level image-to-image translation tasks. In order to capture global structure with local statistics while translation, we design a dual pathway model with one coarse discriminator and one fine discriminator. For generator, we provide one extra perceptual loss in association with adversarial loss and cycle consistency loss to achieve representation learning for two different domains. Also the style can be learned by the auxiliary noise input. Experiments on photo-to-caricature translation of faces in the wild show considerable performance gain of our proposed method over state-of-the-art translation methods as well as its potential real applications.

Chao Wang, Xuanqin Mou, Lei Zhang

In video coding, it is expected that the encoder could adaptively select the encoding parameters (e.g., quantization parameter) to optimize the bit allocation to different sources under the given constraint. However, in hybrid video coding, the dependency between sources brings high complexity for the bit allocation optimization, especially in the block-level, and existing optimization methods mostly focus on frame-level bit allocation. In this paper, we propose a macroblock (MB) level bit allocation method based on the minimum maximum (MINMAX) criterion, which has acceptable encoding complexity for offline applications. An iterative-based algorithm, namely maximum distortion descend (MDD), is developed to reduce quality fluctuation among MBs within a frame, where the Structure SIMilarity (SSIM) index is used to measure the perceptual distortion of MBs. Our extensive experimental results on benchmark video sequences show that the proposed method can greatly enhance the encoding performance in terms of both bits saving and perceptual quality improvement.