Neural Radiance Fields (NeRF) methods have proved effective as compact, high-quality and versatile representations for 3D scenes, and enable downstream tasks such as editing, retrieval, navigation, etc. Various neural architectures are vying for the core structure of NeRF, including the plain Multi-Layer Perceptron (MLP), sparse tensors, low-rank tensors, hashtables and their compositions. Each of these representations has its particular set of trade-offs. For example, the hashtable-based representations admit faster training and rendering but their lack of clear geometric meaning hampers downstream tasks like spatial-relation-aware editing. In this paper, we propose Progressive Volume Distillation (PVD), a systematic distillation method that allows any-to-any conversions between different architectures, including MLP, sparse or low-rank tensors, hashtables and their compositions. PVD consequently empowers downstream applications to optimally adapt the neural representations for the task at hand in a post hoc fashion. The conversions are fast, as distillation is progressively performed on different levels of volume representations, from shallower to deeper. We also employ special treatment of density to deal with its specific numerical instability problem. Empirical evidence is presented to validate our method on the NeRF-Synthetic, LLFF and TanksAndTemples datasets. For example, with PVD, an MLP-based NeRF model can be distilled from a hashtable-based Instant-NGP model at a 10X~20X faster speed than being trained the original NeRF from scratch, while achieving a superior level of synthesis quality. Code is available at https://github.com/megvii-research/AAAI2023-PVD.
Due to the noises in crowdsourced labels, label aggregation (LA) has emerged as a standard procedure to post-process crowdsourced labels. LA methods estimate true labels from crowdsourced labels by modeling worker qualities. Most existing LA methods are iterative in nature. They need to traverse all the crowdsourced labels multiple times in order to jointly and iteratively update true labels and worker qualities until convergence. Consequently, these methods have high space and time complexities. In this paper, we treat LA as a dynamic system and model it as a Dynamic Bayesian network. From the dynamic model we derive two light-weight algorithms, LA\textsuperscript{onepass} and LA\textsuperscript{twopass}, which can effectively and efficiently estimate worker qualities and true labels by traversing all the labels at most twice. Due to the dynamic nature, the proposed algorithms can also estimate true labels online without re-visiting historical data. We theoretically prove the convergence property of the proposed algorithms, and bound the error of estimated worker qualities. We also analyze the space and time complexities of the proposed algorithms and show that they are equivalent to those of majority voting. Experiments conducted on 20 real-world datasets demonstrate that the proposed algorithms can effectively and efficiently aggregate labels in both offline and online settings even if they traverse all the labels at most twice.
Stereo images, containing left and right view images with disparity, are utilized in solving low-vision tasks recently, e.g., rain removal and super-resolution. Stereo image restoration methods usually obtain better performance than monocular methods by learning the disparity between dual views either implicitly or explicitly. However, existing stereo rain removal methods still cannot make full use of the complementary information between two views, and we find it is because: 1) the rain streaks have more complex distributions in directions and densities, which severely damage the complementary information and pose greater challenges; 2) the disparity estimation is not accurate enough due to the imperfect fusion mechanism for the features between two views. To overcome such limitations, we propose a new \underline{Stereo} \underline{I}mage \underline{R}ain \underline{R}emoval method (StereoIRR) via sufficient interaction between two views, which incorporates: 1) a new Dual-view Mutual Attention (DMA) mechanism which generates mutual attention maps by taking left and right views as key information for each other to facilitate cross-view feature fusion; 2) a long-range and cross-view interaction, which is constructed with basic blocks and dual-view mutual attention, can alleviate the adverse effect of rain on complementary information to help the features of stereo images to get long-range and cross-view interaction and fusion. Notably, StereoIRR outperforms other related monocular and stereo image rain removal methods on several datasets. Our codes and datasets will be released.
In this report, we present the ReLER@ZJU1 submission to the Ego4D Moment Queries Challenge in ECCV 2022. In this task, the goal is to retrieve and localize all instances of possible activities in egocentric videos. Ego4D dataset is challenging for the temporal action localization task as the temporal duration of the videos is quite long and each video contains multiple action instances with fine-grained action classes. To address these problems, we utilize a multi-scale transformer to classify different action categories and predict the boundary of each instance. Moreover, in order to better capture the long-term temporal dependencies in the long videos, we propose a segment-level recurrence mechanism. Compared with directly feeding all video features to the transformer encoder, the proposed segment-level recurrence mechanism alleviates the optimization difficulties and achieves better performance. The final submission achieved Recall@1,tIoU=0.5 score of 37.24, average mAP score of 17.67 and took 3-rd place on the leaderboard.
Given the fact description text of a legal case, legal judgment prediction (LJP) aims to predict the case's charge, law article and penalty term. A core problem of LJP is how to distinguish confusing legal cases, where only subtle text differences exist. Previous studies fail to distinguish different classification errors with a standard cross-entropy classification loss, and ignore the numbers in the fact description for predicting the term of penalty. To tackle these issues, in this work, first, we propose a moco-based supervised contrastive learning to learn distinguishable representations, and explore the best strategy to construct positive example pairs to benefit all three subtasks of LJP simultaneously. Second, in order to exploit the numbers in legal cases for predicting the penalty terms of certain cases, we further enhance the representation of the fact description with extracted crime amounts which are encoded by a pre-trained numeracy model. Extensive experiments on public benchmarks show that the proposed method achieves new state-of-the-art results, especially on confusing legal cases. Ablation studies also demonstrate the effectiveness of each component.
Unsupervised Domain Adaptation (UDA) aims to enhance the generalization of the learned model to other domains. The domain-invariant knowledge is transferred from the model trained on labeled source domain, e.g., video game, to unlabeled target domains, e.g., real-world scenarios, saving annotation expenses. Existing UDA methods for semantic segmentation usually focus on minimizing the inter-domain discrepancy of various levels, e.g., pixels, features, and predictions, for extracting domain-invariant knowledge. However, the primary intra-domain knowledge, such as context correlation inside an image, remains underexplored. In an attempt to fill this gap, we propose a unified pixel- and patch-wise self-supervised learning framework, called PiPa, for domain adaptive semantic segmentation that facilitates intra-image pixel-wise correlations and patch-wise semantic consistency against different contexts. The proposed framework exploits the inherent structures of intra-domain images, which: (1) explicitly encourages learning the discriminative pixel-wise features with intra-class compactness and inter-class separability, and (2) motivates the robust feature learning of the identical patch against different contexts or fluctuations. Extensive experiments verify the effectiveness of the proposed method, which obtains competitive accuracy on the two widely-used UDA benchmarks, i.e., 75.6 mIoU on GTA to Cityscapes and 68.2 mIoU on Synthia to Cityscapes. Moreover, our method is compatible with other UDA approaches to further improve the performance without introducing extra parameters.
Low-light stereo image enhancement (LLSIE) is a relatively new task to enhance the quality of visually unpleasant stereo images captured in dark condition. However, current methods achieve inferior performance on detail recovery and illumination adjustment. We find it is because: 1) the insufficient single-scale inter-view interaction makes the cross-view cues unable to be fully exploited; 2) lacking long-range dependency leads to the inability to deal with the spatial long-range effects caused by illumination degradation. To alleviate such limitations, we propose a LLSIE model termed Decoupled Cross-scale Cross-view Interaction Network (DCI-Net). Specifically, we present a decoupled interaction module (DIM) that aims for sufficient dual-view information interaction. DIM decouples the dual-view information exchange into discovering multi-scale cross-view correlations and further exploring cross-scale information flow. Besides, we present a spatial-channel information mining block (SIMB) for intra-view feature extraction, and the benefits are twofold. One is the long-range dependency capture to build spatial long-range relationship, and the other is expanded channel information refinement that enhances information flow in channel dimension. Extensive experiments on Flickr1024, KITTI 2012, KITTI 2015 and Middlebury datasets show that our method obtains better illumination adjustment and detail recovery, and achieves SOTA performance compared to other related methods. Our codes, datasets and models will be publicly available.
This study proposes an improved end-to-end multi-target tracking algorithm that adapts to multi-view multi-scale scenes based on the self-attentive mechanism of the transformer's encoder-decoder structure. A multi-dimensional feature extraction backbone network is combined with a self-built semantic raster map, which is stored in the encoder for correlation and generates target position encoding and multi-dimensional feature vectors. The decoder incorporates four methods: spatial clustering and semantic filtering of multi-view targets, dynamic matching of multi-dimensional features, space-time logic-based multi-target tracking, and space-time convergence network (STCN)-based parameter passing. Through the fusion of multiple decoding methods, muti-camera targets are tracked in three dimensions: temporal logic, spatial logic, and feature matching. For the MOT17 dataset, this study's method significantly outperforms the current state-of-the-art method MiniTrackV2 [49] by 2.2% to 0.836 on Multiple Object Tracking Accuracy(MOTA) metric. Furthermore, this study proposes a retrospective mechanism for the first time, and adopts a reverse-order processing method to optimise the historical mislabeled targets for improving the Identification F1-score(IDF1). For the self-built dataset OVIT-MOT01, the IDF1 improves from 0.948 to 0.967, and the Multi-camera Tracking Accuracy(MCTA) improves from 0.878 to 0.909, which significantly improves the continuous tracking accuracy and scene adaptation. This research method introduces a new attentional tracking paradigm which is able to achieve state-of-the-art performance on multi-target tracking (MOT17 and OVIT-MOT01) tasks.
Cross-view geo-localization aims to spot images of the same location shot from two platforms, e.g., the drone platform and the satellite platform. Existing methods usually focus on optimizing the distance between one embedding with others in the feature space, while neglecting the redundancy of the embedding itself. In this paper, we argue that the low redundancy is also of importance, which motivates the model to mine more diverse patterns. To verify this point, we introduce a simple yet effective regularization, i.e., Dynamic Weighted Decorrelation Regularization (DWDR), to explicitly encourage networks to learn independent embedding channels. As the name implies, DWDR regresses the embedding correlation coefficient matrix to a sparse matrix, i.e., the identity matrix, with dynamic weights. The dynamic weights are applied to focus on still correlated channels during training. Besides, we propose a cross-view symmetric sampling strategy, which keeps the example balance between different platforms. Albeit simple, the proposed method has achieved competitive results on three large-scale benchmarks, i.e., University-1652, CVUSA and CVACT. Moreover, under the harsh circumstance, e.g., the extremely short feature of 64 dimensions, the proposed method surpasses the baseline model by a clear margin.
This paper is about an extraordinary phenomenon. Suppose we don't use any low-light images as training data, can we enhance a low-light image by deep learning? Obviously, current methods cannot do this, since deep neural networks require to train their scads of parameters using copious amounts of training data, especially task-related data. In this paper, we show that in the context of fundamental deep learning, it is possible to enhance a low-light image without any task-related training data. Technically, we propose a new, magical, effective and efficient method, termed \underline{Noi}se \underline{SE}lf-\underline{R}egression (NoiSER), which learns a gray-world mapping from Gaussian distribution for low-light image enhancement (LLIE). Specifically, a self-regression model is built as a carrier to learn a gray-world mapping during training, which is performed by simply iteratively feeding random noise. During inference, a low-light image is directly fed into the learned mapping to yield a normal-light one. Extensive experiments show that our NoiSER is highly competitive to current task-related data based LLIE models in terms of quantitative and visual results, while outperforming them in terms of the number of parameters, training time and inference speed. With only about 1K parameters, NoiSER realizes about 1 minute for training and 1.2 ms for inference with 600$\times$400 resolution on RTX 2080 Ti. Besides, NoiSER has an inborn automated exposure suppression capability and can automatically adjust too bright or too dark, without additional manipulations.