Search

Few-shot action recognition aims to recognize novel action classes (query) using just a few samples (support). The majority of current approaches follow the metric learning paradigm, which learns to compare the similarity between videos. Recently, it has been observed that directly measuring this similarity is not ideal since different action instances may show distinctive temporal distribution, resulting in severe misalignment issues across query and support videos. In this paper, we arrest this problem from two distinct aspects -- action duration misalignment and action evolution misalignment. We address them sequentially through a Two-stage Action Alignment Network (TA2N). The first stage locates the action by learning a temporal affine transform, which warps each video feature to its action duration while dismissing the action-irrelevant feature (e.g. background). Next, the second stage coordinates query feature to match the spatial-temporal action evolution of support by performing temporally rearrange and spatially offset prediction. Extensive experiments on benchmark datasets show the potential of the proposed method in achieving state-of-the-art performance for few-shot action recognition.The code of this project can be found at https://github.com/R00Kie-Liu/TA2N, Published in AAAI 2022

To accelerate DNNs inference, low-rank approximation has been widely adopted because of its solid theoretical rationale and efficient implementations. Several previous works attempted to directly approximate a pre-trained model by low-rank decomposition; however, small approximation errors in parameters can ripple over a large prediction loss. Apparently, it is not optimal to separate low-rank approximation from training. Unlike previous works, this paper integrates low rank approximation and regularization into the training process. We propose Trained Rank Pruning (TRP), which alternates between low rank approximation and training. TRP maintains the capacity of the original network while imposing low-rank constraints during training. A nuclear regularization optimized by stochastic sub-gradient descent is utilized to further promote low rank in TRP. Networks trained with TRP has a low-rank structure in nature, and is approximated with negligible performance loss, thus eliminating fine-tuning after low rank approximation. The proposed method is comprehensively evaluated on CIFAR-10 and ImageNet, outperforming previous compression counterparts using low rank approximation. Our code is available at: https://github.com/yuhuixu1993/Trained-Rank-Pruning., overlap with arXiv:1812.02402, in order to merge the two submissions such that withdraw this version

The proliferation of mobile devices has led to an increasingly interconnected era, and a recent trend is the migration of multimedia search and analysis from the desktop to the cloud. This is, perhaps, partly due to the ease of capturing images and videos on a mobile device and associating them with social and contextual metadata such as comments and GPS tags. Such a hybrid data structure can benefit a wide variety of potential (mobile) multimedia applications, such as location recognition, landmark search, augmented reality, and commercial recommendations. One intrinsic potential solution is to explore large-scale social multimedia to assist and facilitate location related applications on the go and in real-time. With the combination of social and mobile cues, several problems that are previously difficult to be tackled in multimedia content analysis have also become more tractable. Current research in Intelligent Sensing is addressing the following

Depth image super-resolution is a significant yet challenging task. In this paper, we introduce a novel deep color guided coarse-to-fine convolutional neural network (CNN) framework to address this problem. First, we present a data-driven filter method to approximate the ideal filter for depth image super-resolution instead of hand-designed filters. Based on large data samples, the filter learned is more accurate and stable for upsampling depth image. Second, we introduce a coarse-to-fine CNN to learn different sizes of filter kernels. In the coarse stage, larger filter kernels are learned by the CNN to achieve crude high-resolution depth image. As to the fine stage, the crude high-resolution depth image is used as the input so that smaller filter kernels are learned to gain more accurate results. Benefit from this network, we can progressively recover the high frequency details. Third, we construct a color guidance strategy that fuses color difference and spatial distance for depth image upsampling. We revise the interpolated high-resolution depth image according to the corresponding pixels in high-resolution color maps. Guided by color information, the depth of high-resolution image obtained can alleviate texture copying artifacts and preserve edge details effectively. Quantitative and qualitative experimental results demonstrate our state-of-the-art performance for depth map super-resolution.

Object detection has made great progress in the past few years along with the development of deep learning. However, most current object detection methods are resource hungry, which hinders their wide deployment to many resource restricted usages such as usages on always-on devices, battery-powered low-end devices, etc. This paper considers the resource and accuracy trade-off for resource-restricted usages during designing the whole object detection framework. Based on the deeply supervised object detection (DSOD) framework, we propose Tiny-DSOD dedicating to resource-restricted usages. Tiny-DSOD introduces two innovative and ultra-efficient architecture blocks: depthwise dense block (DDB) based backbone and depthwise feature-pyramid-network (D-FPN) based front-end. We conduct extensive experiments on three famous benchmarks (PASCAL VOC 2007, KITTI, and COCO), and compare Tiny-DSOD to the state-of-the-art ultra-efficient object detection solutions such as Tiny-YOLO, MobileNet-SSD (v1 & v2), SqueezeDet, Pelee, etc. Results show that Tiny-DSOD outperforms these solutions in all the three metrics (parameter-size, FLOPs, accuracy) in each comparison. For instance, Tiny-DSOD achieves 72.1% mAP with only 0.95M parameters and 1.06B FLOPs, which is by far the state-of-the-arts result with such a low resource requirement., 12 pages, 3 figures, accepted by BMVC 2018

In this paper, we propose a partition-masked Convolution Neural Network (CNN) to achieve compressed-video enhancement for the state-of-the-art coding standard, High Efficiency Video Coding (HECV). More precisely, our method utilizes the partition information produced by the encoder to guide the quality enhancement process. In contrast to existing CNN-based approaches, which only take the decoded frame as the input to the CNN, the proposed approach considers the coding unit (CU) size information and combines it with the distorted decoded frame such that the degradation introduced by HEVC is reduced more efficiently. Experimental results show that our approach leads to over 9.76% BD-rate saving on benchmark sequences, which achieves the state-of-the-art performance., Accepted by ICIP 2018

This paper addresses the problem of handling spatial misalignments due to camera-view changes or human-pose variations in person re-identification. We first introduce a boosting-based approach to learn a correspondence structure which indicates the patch-wise matching probabilities between images from a target camera pair. The learned correspondence structure can not only capture the spatial correspondence pattern between cameras but also handle the viewpoint or human-pose variation in individual images. We further introduce a global constraint-based matching process. It integrates a global matching constraint over the learned correspondence structure to exclude cross-view misalignments during the image patch matching process, hence achieving a more reliable matching score between images. Finally, we also extend our approach by introducing a multi-structure scheme, which learns a set of local correspondence structures to capture the spatial correspondence sub-patterns between a camera pair, so as to handle the spatial misalignments between individual images in a more precise way. Experimental results on various datasets demonstrate the effectiveness of our approach., IEEE Trans. Image Processing, vol. 26, no. 5, pp. 2438-2453, 2017. The project page for this paper is available at http://min.sjtu.edu.cn/lwydemo/personReID.htm arXiv admin note: text overlap with arXiv:1504.06243

This paper presents a novel representative-based framework for parsing and summarizing events in long surveillance videos. The proposed framework first extracts object blob sequences and utilizes them to represent events in a surveillance video. Then, a sequence filtering strategy is introduced which detects and eliminates noisy blob sequences based on their spatial and temporal characteristics. After clustering the blob sequences into different event types, we further introduce a representative-based model which integrates location, size, and appearance cues to select a representative blob sequence from each cluster, and creates a snapshot image for each representative blob sequence. Based on the blob-sequence clustering and representative-sequence selection results, two schemes are further proposed to summarize contents of the input surveillance video: (1) type-based scheme which shows snapshot images to users and creates a summary video for a specific event cluster according to user-selected snapshot image; (2) representative-based scheme which creates a summary video only with the extracted representative blob sequences. Experimental results show that our approach can create more effective and well-organized summarization results compared with the state-of-the-art methods.

The visualization of an image collection is the process of displaying a collection of images on a screen under some specific layout requirements. This paper focuses on an important problem that is not well addressed by the previous methods: visualizing image collections into arbitrary layout shapes while arranging images according to user-defined semantic or visual correlations (e.g., color or object category). To this end, we first propose a property-based tree construction scheme to organize images of a collection into a tree structure according to user-defined properties. In this way, images can be adaptively placed with the desired semantic or visual correlations in the final visualization layout. Then, we design a two-step visualization optimization scheme to further optimize image layouts. As a result, multiple layout effects including layout shape and image overlap ratio can be effectively controlled to guarantee a satisfactory visualization. Finally, we also propose a tree-transfer scheme such that visualization layouts can be adaptively changed when users select different "images of interest". We demonstrate the effectiveness of our proposed approach through the comparisons with state-of-the-art visualization techniques., This manuscript is the accepted version for T-CYB (IEEE Transactions on Cybernetics) IEEE Trans. Cybernetics, 2015

This paper addresses the problem of handling spatial misalignments due to camera-view changes or human-pose variations in person re-identification. We first introduce a boosting-based approach to learn a correspondence structure which indicates the patch-wise matching probabilities between images from a target camera pair. The learned correspondence structure can not only capture the spatial correspondence pattern between cameras but also handle the viewpoint or human-pose variation in individual images. We further introduce a global-based matching process. It integrates a global matching constraint over the learned correspondence structure to exclude cross-view misalignments during the image patch matching process, hence achieving a more reliable matching score between images. Experimental results on various datasets demonstrate the effectiveness of our approach.

This paper presents a novel approach for automatic recognition of human activities for video surveillance applications. We propose to represent an activity by a combination of category components, and demonstrate that this approach offers flexibility to add new activities to the system and an ability to deal with the problem of building models for activities lacking training data. For improving the recognition accuracy, a Confident-Frame- based Recognition algorithm is also proposed, where the video frames with high confidence for recognizing an activity are used as a specialized local model to help classify the remainder of the video frames. Experimental results show the effectiveness of the proposed approach., This manuscript is the accepted version for TCSVT (IEEE Transactions on Circuits and Systems for Video Technology)

In this paper, a macroblock classification method is proposed for various video processing applications involving motions. Based on the analysis of the Motion Vector field in the compressed video, we propose to classify Macroblocks of each video frame into different classes and use this class information to describe the frame content. We demonstrate that this low-computation-complexity method can efficiently catch the characteristics of the frame. Based on the proposed macroblock classification, we further propose algorithms for different video processing applications, including shot change detection, motion discontinuity detection, and outlier rejection for global motion estimation. Experimental results demonstrate that the methods based on the proposed approach can work effectively on these applications., This manuscript is the accepted version for TB (IEEE Transactions on Broadcasting)

The Visual Object Tracking challenge 2014, VOT2014, aims at comparing short-term single-object visual trackers that do not apply pre-learned models of object appearance. Results of 38 trackers are presented. The number of tested trackers makes VOT 2014 the largest benchmark on short-term tracking to date. For each participating tracker, a short description is provided in the appendix. Features of the VOT2014 challenge that go beyond its VOT2013 predecessor are introduced: (i) a new VOT2014 dataset with full annotation of targets by rotated bounding boxes and per-frame attribute, (ii) extensions of the VOT2013 evaluation methodology, (iii) a new unit for tracking speed assessment less dependent on the hardware and (iv) the VOT2014 evaluation toolkit that significantly speeds up execution of experiments. The dataset, the evaluation kit as well as the results are publicly available at the challenge website (http://​votchallenge.​net).

International audience; The most popular applications for video understanding are those related to video-surveillance (e.g. alarms, abnormal behaviours, expected events, access control). Video understanding has several other applications of high impact to the society as medical supervision, traffic control, violent acts detection, crowd behaviour analysis, among many others. We propose a new generic video understanding approach able to extract and learn valuable information from noisy video scenes for real-time applications. This approach comprises motion segmentation, object classification, tracking and event learning phases. This work is focused on building the first fundamental blocks allowing a proper management of uncertainty of data in every phase of the video understanding process. The main contributions of the proposed approach are: (i) a new algorithm for tracking multiple objects in noisy environments, (ii) the utilisation of reliability measures for modelling uncertainty in data and for proper selection of valuable information extracted from noisy data, (iii) the improved capability of tracking to manage multiple visual evidence-target associations, (iv) the combination of 2D image data with 3D information in a dynamics model governed by reliability measures for proper control of uncertainty in data, and (v) a new approach for event recognition through incremental event learning, driven by reliability measures for selecting the most stable and relevant data.