In dynamic scenes, images often suffer from dynamic blur due to superposition of motions or low signal-noise ratio resulted from quick shutter speed when avoiding motions. Recovering sharp and clean results from the captured images heavily depends on the ability of restoration methods and the quality of the input. Although existing research on image restoration focuses on developing models for obtaining better restored results, fewer have studied to evaluate how and which input image leads to superior restored quality. In this paper, to better study an image's potential value that can be explored for restoration, we propose a novel concept, referring to image restoration potential (IRP). Specifically, We first establish a dynamic scene imaging dataset containing composite distortions and applied image restoration processes to validate the rationality of the existence to IRP. Based on this dataset, we investigate several properties of IRP and propose a novel deep model to accurately predict IRP values. By gradually distilling and selective fusing the degradation features, the proposed model shows its superiority in IRP prediction. Thanks to the proposed model, we are then able to validate how various image restoration related applications are benefited from IRP prediction. We show the potential usages of IRP as a filtering principle to select valuable frames, an auxiliary guidance to improve restoration models, and even an indicator to optimize camera settings for capturing better images under dynamic scenarios.
DeepFake based digital facial forgery is threatening the public media security, especially when lip manipulation has been used in talking face generation, the difficulty of fake video detection is further improved. By only changing lip shape to match the given speech, the facial features of identity is hard to be discriminated in such fake talking face videos. Together with the lack of attention on audio stream as the prior knowledge, the detection failure of fake talking face generation also becomes inevitable. Inspired by the decision-making mechanism of human multisensory perception system, which enables the auditory information to enhance post-sensory visual evidence for informed decisions output, in this study, a fake talking face detection framework FTFDNet is proposed by incorporating audio and visual representation to achieve more accurate fake talking face videos detection. Furthermore, an audio-visual attention mechanism (AVAM) is proposed to discover more informative features, which can be seamlessly integrated into any audio-visual CNN architectures by modularization. With the additional AVAM, the proposed FTFDNet is able to achieve a better detection performance on the established dataset (FTFDD). The evaluation of the proposed work has shown an excellent performance on the detection of fake talking face videos, which is able to arrive at a detection rate above 97%.
Multi-modal based speech separation has exhibited a specific advantage on isolating the target character in multi-talker noisy environments. Unfortunately, most of current separation strategies prefer a straightforward fusion based on feature learning of each single modality, which is far from sufficient consideration of inter-relationships between modalites. Inspired by learning joint feature representations from audio and visual streams with attention mechanism, in this study, a novel cross-modal fusion strategy is proposed to benefit the whole framework with semantic correlations between different modalities. To further improve audio-visual speech separation, the dense optical flow of lip motion is incorporated to strengthen the robustness of visual representation. The evaluation of the proposed work is performed on two public audio-visual speech separation benchmark datasets. The overall improvement of the performance has demonstrated that the additional motion network effectively enhances the visual representation of the combined lip images and audio signal, as well as outperforming the baseline in terms of all metrics with the proposed cross-modal fusion.
For the weakly supervised anomaly detection task, most existing work is limited to the problem of inadequate video representation due to the inability to model long-time contextual information. We propose a weakly supervised adaptive graph convolutional network (WAGCN) to model the contextual relationships among video segments. And we fully consider the influence of other video segments on the current segment when generating the anomaly probability score for each segment. Firstly, we combine the temporal consistency as well as feature similarity of video segments for composition, which makes full use of the association information among spatial-temporal features of anomalous events in videos. Secondly, we propose a graph learning layer in order to break the limitation of setting topology manually, which adaptively extracts sparse graph adjacency matrix based on data. Extensive experiments on two public datasets (i.e., UCF-Crime dataset and ShanghaiTech dataset) demonstrate the effectiveness of our approach.
Deep learning-based person Re-IDentification (ReID) often requires a large amount of training data to achieve good performance. Thus it appears that collecting more training data from diverse environments tends to improve the ReID performance. This paper re-examines this common belief and makes a somehow surprising observation: using more samples, i.e., training with samples from multiple datasets, does not necessarily lead to better performance by using the popular ReID models. In some cases, training with more samples may even hurt the performance of the evaluation is carried out in one of those datasets. We postulate that this phenomenon is due to the incapability of the standard network in adapting to diverse environments. To overcome this issue, we propose an approach called Domain-Camera-Sample Dynamic network (DCSD) whose parameters can be adaptive to various factors. Specifically, we consider the internal domain-related factor that can be identified from the input features, and external domain-related factors, such as domain information or camera information. Our discovery is that training with such an adaptive model can better benefit from more training samples. Experimental results show that our DCSD can greatly boost the performance (up to 12.3%) while joint training in multiple datasets.
Video anomaly detection (VAD) mainly refers to identifying anomalous events that have not occurred in the training set where only normal samples are available. Existing works usually formulate VAD as a reconstruction or prediction problem. However, the adaptability and scalability of these methods are limited. In this paper, we propose a novel distance-based VAD method to take advantage of all the available normal data efficiently and flexibly. In our method, the smaller the distance between a testing sample and normal samples, the higher the probability that the testing sample is normal. Specifically, we propose to use locality-sensitive hashing (LSH) to map samples whose similarity exceeds a certain threshold into the same bucket in advance. In this manner, the complexity of near neighbor search is cut down significantly. To make the samples that are semantically similar get closer and samples not similar get further apart, we propose a novel learnable version of LSH that embeds LSH into a neural network and optimizes the hash functions with contrastive learning strategy. The proposed method is robust to data imbalance and can handle the large intra-class variations in normal data flexibly. Besides, it has a good ability of scalability. Extensive experiments demonstrate the superiority of our method, which achieves new state-of-the-art results on VAD benchmarks.
Neural Architecture Search (NAS) has shown great potential in effectively reducing manual effort in network design by automatically discovering optimal architectures. What is noteworthy is that as of now, object detection is less touched by NAS algorithms despite its significant importance in computer vision. To the best of our knowledge, most of the recent NAS studies on object detection tasks fail to satisfactorily strike a balance between performance and efficiency of the resulting models, let alone the excessive amount of computational resources cost by those algorithms. Here we propose an efficient method to obtain better object detectors by searching for the feature pyramid network (FPN) as well as the prediction head of a simple anchor-free object detector, namely, FCOS [36], using a tailored reinforcement learning paradigm. With carefully designed search space, search algorithms, and strategies for evaluating network quality, we are able to find top-performing detection architectures within 4 days using 8 V100 GPUs. The discovered architectures surpass state-of-the-art object detection models (such as Faster R-CNN, Retina-Net and, FCOS) by 1.0% to 5.4% points in AP on the COCO dataset, with comparable computation complexity and memory footprint, demonstrating the efficacy of the proposed NAS method for object detection. Code is available at https://github.com/Lausannen/NAS-FCOS.
Finding target persons in full scene images with a query of text description has important practical applications in intelligent video surveillance.However, different from the real-world scenarios where the bounding boxes are not available, existing text-based person retrieval methods mainly focus on the cross modal matching between the query text descriptions and the gallery of cropped pedestrian images. To close the gap, we study the problem of text-based person search in full images by proposing a new end-to-end learning framework which jointly optimize the pedestrian detection, identification and visual-semantic feature embedding tasks. To take full advantage of the query text, the semantic features are leveraged to instruct the Region Proposal Network to pay more attention to the text-described proposals. Besides, a cross-scale visual-semantic embedding mechanism is utilized to improve the performance. To validate the proposed method, we collect and annotate two large-scale benchmark datasets based on the widely adopted image-based person search datasets CUHK-SYSU and PRW. Comprehensive experiments are conducted on the two datasets and compared with the baseline methods, our method achieves the state-of-the-art performance.
The target representation learned by convolutional neural networks plays an important role in Thermal Infrared (TIR) tracking. Currently, most of the top-performing TIR trackers are still employing representations learned by the model trained on the RGB data. However, this representation does not take into account the information in the TIR modality itself, limiting the performance of TIR tracking. To solve this problem, we propose to distill representations of the TIR modality from the RGB modality with Cross-Modal Distillation (CMD) on a large amount of unlabeled paired RGB-TIR data. We take advantage of the two-branch architecture of the baseline tracker, i.e. DiMP, for cross-modal distillation working on two components of the tracker. Specifically, we use one branch as a teacher module to distill the representation learned by the model into the other branch. Benefiting from the powerful model in the RGB modality, the cross-modal distillation can learn the TIR-specific representation for promoting TIR tracking. The proposed approach can be incorporated into different baseline trackers conveniently as a generic and independent component. Furthermore, the semantic coherence of paired RGB and TIR images is utilized as a supervised signal in the distillation loss for cross-modal knowledge transfer. In practice, three different approaches are explored to generate paired RGB-TIR patches with the same semantics for training in an unsupervised way. It is easy to extend to an even larger scale of unlabeled training data. Extensive experiments on the LSOTB-TIR dataset and PTB-TIR dataset demonstrate that our proposed cross-modal distillation method effectively learns TIR-specific target representations transferred from the RGB modality. Our tracker outperforms the baseline tracker by achieving absolute gains of 2.3% Success, 2.7% Precision, and 2.5% Normalized Precision respectively.
With the explosive growth of video data, video summarization, which attempts to seek the minimum subset of frames while still conveying the main story, has become one of the hottest topics. Nowadays, substantial achievements have been made by supervised learning techniques, especially after the emergence of deep learning. However, it is extremely expensive and difficult to collect human annotation for large-scale video datasets. To address this problem, we propose a convolutional attentive adversarial network (CAAN), whose key idea is to build a deep summarizer in an unsupervised way. Upon the generative adversarial network, our overall framework consists of a generator and a discriminator. The former predicts importance scores for all frames of a video while the latter tries to distinguish the score-weighted frame features from original frame features. Specifically, the generator employs a fully convolutional sequence network to extract global representation of a video, and an attention-based network to output normalized importance scores. To learn the parameters, our objective function is composed of three loss functions, which can guide the frame-level importance score prediction collaboratively. To validate this proposed method, we have conducted extensive experiments on two public benchmarks SumMe and TVSum. The results show the superiority of our proposed method against other state-of-the-art unsupervised approaches. Our method even outperforms some published supervised approaches.