The deep-learning-based video coding has attracted substantial attention for its great potential to squeeze out the spatial-temporal redundancies of video sequences. This paper proposes an efficient codec namely dual-path generative adversarial network-based video codec (DGVC). First, we propose a dual-path enhancement with generative adversarial network (DPEG) to reconstruct the compressed video details. The DPEG consists of an $\alpha$-path of auto-encoder and convolutional long short-term memory (ConvLSTM), which facilitates the structure feature reconstruction with a large receptive field and multi-frame references, and a $\beta$-path of residual attention blocks, which facilitates the reconstruction of local texture features. Both paths are fused and co-trained by a generative-adversarial process. Second, we reuse the DPEG network in both motion compensation and quality enhancement modules, which are further combined with motion estimation and entropy coding modules in our DGVC framework. Third, we employ a joint training of deep video compression and enhancement to further improve the rate-distortion (RD) performance. Compared with x265 LDP very fast mode, our DGVC reduces the average bit-per-pixel (bpp) by 39.39%/54.92% at the same PSNR/MS-SSIM, which outperforms the state-of-the art deep video codecs by a considerable margin.
Video Coding for Machines (VCM) is committed to bridging to an extent separate research tracks of video/image compression and feature compression, and attempts to optimize compactness and efficiency jointly from a unified perspective of high accuracy machine vision and full fidelity human vision. In this paper, we summarize VCM methodology and philosophy based on existing academia and industrial efforts. The development of VCM follows a general rate-distortion optimization, and the categorization of key modules or techniques is established. From previous works, it is demonstrated that, although existing works attempt to reveal the nature of scalable representation in bits when dealing with machine and human vision tasks, there remains a rare study in the generality of low bit rate representation, and accordingly how to support a variety of visual analytic tasks. Therefore, we investigate a novel visual information compression for the analytics taxonomy problem to strengthen the capability of compact visual representations extracted from multiple tasks for visual analytics. A new perspective of task relationships versus compression is revisited. By keeping in mind the transferability among different machine vision tasks (e.g. high-level semantic and mid-level geometry-related), we aim to support multiple tasks jointly at low bit rates. In particular, to narrow the dimensionality gap between neural network generated features extracted from pixels and a variety of machine vision features/labels (e.g. scene class, segmentation labels), a codebook hyperprior is designed to compress the neural network-generated features. As demonstrated in our experiments, this new hyperprior model is expected to improve feature compression efficiency by estimating the signal entropy more accurately, which enables further investigation of the granularity of abstracting compact features among different tasks.
Self-tracking can improve people's awareness of their unhealthy behaviors to provide insights towards behavior change. Prior work has explored how self-trackers reflect on their logged data, but it remains unclear how much they learn from the tracking feedback, and which information is more useful. Indeed, the feedback can still be overwhelming, and making it concise can improve learning by increasing focus and reducing interpretation burden. We conducted a field study of mobile food logging with two feedback modes (manual journaling and automatic annotation of food images) and identified learning differences regarding nutrition, assessment, behavioral, and contextual information. We propose a Self-Tracking Feedback Saliency Framework to define when to provide feedback, on which specific information, why those details, and how to present them (as manual inquiry or automatic feedback). We propose SalienTrack to implement these requirements. Using the data collected from the user study, we trained a machine learning model to predict whether a user would learn from each tracked event. Using explainable AI (XAI) techniques, we identified the most salient features per instance and why they lead to positive learning outcomes. We discuss implications for learnability in self-tracking, and how adding model explainability expands opportunities for improving feedback experience.
Visual analytics have played an increasingly critical role in the Internet of Things, where massive visual signals have to be compressed and fed into machines. But facing such big data and constrained bandwidth capacity, existing image/video compression methods lead to very low-quality representations, while existing feature compression techniques fail to support diversified visual analytics applications/tasks with low-bit-rate representations. In this paper, we raise and study the novel problem of supporting multiple machine vision analytics tasks with the compressed visual representation, namely, the information compression problem in analytics taxonomy. By utilizing the intrinsic transferability among different tasks, our framework successfully constructs compact and expressive representations at low bit-rates to support a diversified set of machine vision tasks, including both high-level semantic-related tasks and mid-level geometry analytic tasks. In order to impose compactness in the representations, we propose a codebook-based hyperprior, which helps map the representation into a low-dimensional manifold. As it well fits the signal structure of the deep visual feature, it facilitates more accurate entropy estimation, and results in higher compression efficiency. With the proposed framework and the codebook-based hyperprior, we further investigate the relationship of different task features owning different levels of abstraction granularity. Experimental results demonstrate that with the proposed scheme, a set of diversified tasks can be supported at a significantly lower bit-rate, compared with existing compression schemes.
Vision Transformer (ViT) attains state-of-the-art performance in visual recognition, and the variant, Local Vision Transformer, makes further improvements. The major component in Local Vision Transformer, local attention, performs the attention separately over small local windows. We rephrase local attention as a channel-wise locally-connected layer and analyze it from two network regularization manners, sparse connectivity and weight sharing, as well as weight computation. Sparse connectivity: there is no connection across channels, and each position is connected to the positions within a small local window. Weight sharing: the connection weights for one position are shared across channels or within each group of channels. Dynamic weight: the connection weights are dynamically predicted according to each image instance. We point out that local attention resembles depth-wise convolution and its dynamic version in sparse connectivity. The main difference lies in weight sharing - depth-wise convolution shares connection weights (kernel weights) across spatial positions. We empirically observe that the models based on depth-wise convolution and the dynamic variant with lower computation complexity perform on-par with or sometimes slightly better than Swin Transformer, an instance of Local Vision Transformer, for ImageNet classification, COCO object detection and ADE semantic segmentation. These observations suggest that Local Vision Transformer takes advantage of two regularization forms and dynamic weight to increase the network capacity.
Face detection in low light scenarios is challenging but vital to many practical applications, e.g., surveillance video, autonomous driving at night. Most existing face detectors heavily rely on extensive annotations, while collecting data is time-consuming and laborious. To reduce the burden of building new datasets for low light conditions, we make full use of existing normal light data and explore how to adapt face detectors from normal light to low light. The challenge of this task is that the gap between normal and low light is too huge and complex for both pixel-level and object-level. Therefore, most existing low-light enhancement and adaptation methods do not achieve desirable performance. To address the issue, we propose a joint High-Low Adaptation (HLA) framework. Through a bidirectional low-level adaptation and multi-task high-level adaptation scheme, our HLA-Face outperforms state-of-the-art methods even without using dark face labels for training. Our project is publicly available at https://daooshee.github.io/HLA-Face-Website/
In this paper, we address the problem of referring expression comprehension in videos, which is challenging due to complex expression and scene dynamics. Unlike previous methods which solve the problem in multiple stages (i.e., tracking, proposal-based matching), we tackle the problem from a novel perspective, \textbf{co-grounding}, with an elegant one-stage framework. We enhance the single-frame grounding accuracy by semantic attention learning and improve the cross-frame grounding consistency with co-grounding feature learning. Semantic attention learning explicitly parses referring cues in different attributes to reduce the ambiguity in the complex expression. Co-grounding feature learning boosts visual feature representations by integrating temporal correlation to reduce the ambiguity caused by scene dynamics. Experiment results demonstrate the superiority of our framework on the video grounding datasets VID and LiOTB in generating accurate and stable results across frames. Our model is also applicable to referring expression comprehension in images, illustrated by the improved performance on the RefCOCO dataset. Our project is available at https://sijiesong.github.io/co-grounding.
The recent physical model-free dehazing methods have achieved state-of-the-art performances. However, without the guidance of physical models, the performances degrade rapidly when applied to real scenarios due to the unavailable or insufficient data problems. On the other hand, the physical model-based methods have better interpretability but suffer from multi-objective optimizations of parameters, which may lead to sub-optimal dehazing results. In this paper, a progressive residual learning strategy has been proposed to combine the physical model-free dehazing process with reformulated scattering model-based dehazing operations, which enjoys the merits of dehazing methods in both categories. Specifically, the global atmosphere light and transmission maps are interactively optimized with the aid of accurate residual information and preliminary dehazed restorations from the initial physical model-free dehazing process. The proposed method performs favorably against the state-of-the-art methods on public dehazing benchmarks with better model interpretability and adaptivity for complex hazy data.
Features representation leverages the great power in network analysis tasks. However, most features are discrete which poses tremendous challenges to effective use. Recently, increasing attention has been paid on network feature learning, which could map discrete features to continued space. Unfortunately, current studies fail to fully preserve the structural information in the feature space due to random negative sampling strategy during training. To tackle this problem, we study the problem of feature learning and novelty propose a force-based graph learning model named GForce inspired by the spring-electrical model. GForce assumes that nodes are in attractive forces and repulsive forces, thus leading to the same representation with the original structural information in feature learning. Comprehensive experiments on benchmark datasets demonstrate the effectiveness of the proposed framework. Furthermore, GForce opens up opportunities to use physics models to model node interaction for graph learning.
The formulation of the hazy image is mainly dominated by the reflected lights and ambient airlight. Existing dehazing methods often ignore the depth cues and fail in distant areas where heavier haze disturbs the visibility. However, we note that the guidance of the depth information for transmission estimation could remedy the decreased visibility as distances increase. In turn, the good transmission estimation could facilitate the depth estimation for hazy images. In this paper, a deep end-to-end model that iteratively estimates image depths and transmission maps is proposed to perform an effective depth prediction for hazy images and improve the dehazing performance with the guidance of depth information. The image depth and transmission map are progressively refined to better restore the dehazed image. Our approach benefits from explicitly modeling the inner relationship of image depth and transmission map, which is especially effective for distant hazy areas. Extensive results on the benchmarks demonstrate that our proposed network performs favorably against the state-of-the-art dehazing methods in terms of depth estimation and haze removal.