Convolutional networks reach top quality in pixel-level video object segmentation but require a large amount of training data (1k~100k) to deliver such results. We propose a new training strategy which achieves state-of-the-art results across three evaluation datasets while using 20x~1000x less annotated data than competing methods. Our approach is suitable for both single and multiple object segmentation. Instead of using large training sets hoping to generalize across domains, we generate in-domain training data using the provided annotation on the first frame of each video to synthesize ("lucid dream") plausible future video frames. In-domain per-video training data allows us to train high quality appearance- and motion-based models, as well as tune the post-processing stage. This approach allows to reach competitive results even when training from only a single annotated frame, without ImageNet pre-training. Our results indicate that using a larger training set is not automatically better, and that for the video object segmentation task a smaller training set that is closer to the target domain is more effective. This changes the mindset regarding how many training samples and general "objectness" knowledge are required for the video object segmentation task.
Video image datasets are playing an essential role in design and evaluation of traffic vision algorithms. Nevertheless, a longstanding inconvenience concerning image datasets is that manually collecting and annotating large-scale diversified datasets from real scenes is time-consuming and prone to error. For that virtual datasets have begun to function as a proxy of real datasets. In this paper, we propose to construct large-scale artificial scenes for traffic vision research and generate a new virtual dataset called "ParallelEye". First of all, the street map data is used to build 3D scene model of Zhongguancun Area, Beijing. Then, the computer graphics, virtual reality, and rule modeling technologies are utilized to synthesize large-scale, realistic virtual urban traffic scenes, in which the fidelity and geography match the real world well. Furthermore, the Unity3D platform is used to render the artificial scenes and generate accurate ground-truth labels, e.g., semantic/instance segmentation, object bounding box, object tracking, optical flow, and depth. The environmental conditions in artificial scenes can be controlled completely. As a result, we present a viable implementation pipeline for constructing large-scale artificial scenes for traffic vision research. The experimental results demonstrate that this pipeline is able to generate photorealistic virtual datasets with low modeling time and high accuracy labeling.
From human crowds to cells in tissue, the detection and efficient tracking of multiple objects in dense configurations is an important and unsolved problem. In the past, limitations of image analysis have restricted studies of dense groups to tracking a single or subset of marked individuals, or to coarse-grained group-level dynamics, all of which yield incomplete information. Here, we combine convolutional neural networks (CNNs) with the model environment of a honeybee hive to automatically recognize all individuals in a dense group from raw image data. We create new, adapted individual labeling and use the segmentation architecture U-Net with a loss function dependent on both object identity and orientation. We additionally exploit temporal regularities of the video recording in a recurrent manner and achieve near human-level performance while reducing the network size by 94% compared to the original U-Net architecture. Given our novel application of CNNs, we generate extensive problem-specific image data in which labeled examples are produced through a custom interface with Amazon Mechanical Turk. This dataset contains over 375,000 labeled bee instances across 720 video frames at 2 FPS, representing an extensive resource for the development and testing of tracking methods. We correctly detect 96% of individuals with a location error of ~7% of a typical body dimension, and orientation error of 12 degrees, approximating the variability of human raters. Our results provide an important step towards efficient image-based dense object tracking by allowing for the accurate determination of object location and orientation across time-series image data efficiently within one network architecture.
The most common paradigm for vision-based multi-object tracking is tracking-by-detection, due to the availability of reliable detectors for several important object categories such as cars and pedestrians. However, future mobile systems will need a capability to cope with rich human-made environments, in which obtaining detectors for every possible object category would be infeasible. In this paper, we propose a model-free multi-object tracking approach that uses a category-agnostic image segmentation method to track objects. We present an efficient segmentation mask-based tracker which associates pixel-precise masks reported by the segmentation. Our approach can utilize semantic information whenever it is available for classifying objects at the track level, while retaining the capability to track generic unknown objects in the absence of such information. We demonstrate experimentally that our approach achieves performance comparable to state-of-the-art tracking-by-detection methods for popular object categories such as cars and pedestrians. Additionally, we show that the proposed method can discover and robustly track a large variety of other objects.
We explore object discovery and detector adaptation based on unlabeled video sequences captured from a mobile platform. We propose a fully automatic approach for object mining from video which builds upon a generic object tracking approach. By applying this method to three large video datasets from autonomous driving and mobile robotics scenarios, we demonstrate its robustness and generality. Based on the object mining results, we propose a novel approach for unsupervised object discovery by appearance-based clustering. We show that this approach successfully discovers interesting objects relevant to driving scenarios. In addition, we perform self-supervised detector adaptation in order to improve detection performance on the KITTI dataset for existing categories. Our approach has direct relevance for enabling large-scale object learning for autonomous driving.
Current convolutional neural networks algorithms for video object tracking spend the same amount of computation for each object and video frame. However, it is harder to track an object in some frames than others, due to the varying amount of clutter, scene complexity, amount of motion, and object's distinctiveness against its background. We propose a depth-adaptive convolutional Siamese network that performs video tracking adaptively at multiple neural network depths. Parametric gating functions are trained to control the depth of the convolutional feature extractor by minimizing a joint loss of computational cost and tracking error. Our network achieves accuracy comparable to the state-of-the-art on the VOT2016 benchmark. Furthermore, our adaptive depth computation achieves higher accuracy for a given computational cost than traditional fixed-structure neural networks. The presented framework extends to other tasks that use convolutional neural networks and enables trading speed for accuracy at runtime.
Recently, deep learning has achieved very promising results in visual object tracking. Deep neural networks in existing tracking methods require a lot of training data to learn a large number of parameters. However, training data is not sufficient for visual object tracking as annotations of a target object are only available in the first frame of a test sequence. In this paper, we propose to learn hierarchical features for visual object tracking by using tree structure based Recursive Neural Networks (RNN), which have fewer parameters than other deep neural networks, e.g. Convolutional Neural Networks (CNN). First, we learn RNN parameters to discriminate between the target object and background in the first frame of a test sequence. Tree structure over local patches of an exemplar region is randomly generated by using a bottom-up greedy search strategy. Given the learned RNN parameters, we create two dictionaries regarding target regions and corresponding local patches based on the learned hierarchical features from both top and leaf nodes of multiple random trees. In each of the subsequent frames, we conduct sparse dictionary coding on all candidates to select the best candidate as the new target location. In addition, we online update two dictionaries to handle appearance changes of target objects. Experimental results demonstrate that our feature learning algorithm can significantly improve tracking performance on benchmark datasets.
Computer vision technologies are very attractive for practical applications running on embedded systems. For such an application, it is desirable for the deployed algorithms to run in high-speed and require no offline training. To develop a single-target tracking algorithm with these properties, we propose an ensemble of the kernelized correlation filters (KCF), we call it EnKCF. A committee of KCFs is specifically designed to address the variations in scale and translation of moving objects. To guarantee a high-speed run-time performance, we deploy each of KCFs in turn, instead of applying multiple KCFs to each frame. To minimize any potential drifts between individual KCFs transition, we developed a particle filter. Experimental results showed that the performance of ours is, on average, 70.10% for precision at 20 pixels, 53.00% for success rate for the OTB100 data, and 54.50% and 40.2% for the UAV123 data. Experimental results showed that our method is better than other high-speed trackers over 5% on precision on 20 pixels and 10-20% on AUC on average. Moreover, our implementation ran at 340 fps for the OTB100 and at 416 fps for the UAV123 dataset that is faster than DCF (292 fps) for the OTB100 and KCF (292 fps) for the UAV123. To increase flexibility of the proposed EnKCF running on various platforms, we also explored different levels of deep convolutional features.
This work introduces a novel framework for quantifying the presence and strength of recurrent dynamics in video data. Specifically, we provide continuous measures of periodicity (perfect repetition) and quasiperiodicity (superposition of periodic modes with non-commensurate periods), in a way which does not require segmentation, training, object tracking or 1-dimensional surrogate signals. Our methodology operates directly on video data. The approach combines ideas from nonlinear time series analysis (delay embeddings) and computational topology (persistent homology), by translating the problem of finding recurrent dynamics in video data, into the problem of determining the circularity or toroidality of an associated geometric space. Through extensive testing, we show the robustness of our scores with respect to several noise models/levels, we show that our periodicity score is superior to other methods when compared to human-generated periodicity rankings, and furthermore, we show that our quasiperiodicity score clearly indicates the presence of biphonation in videos of vibrating vocal folds, which has never before been accomplished end to end quantitatively.
Being intensively studied, visual object tracking has witnessed great advances in either speed (e.g., with correlation filters) or accuracy (e.g., with deep features). Real-time and high accuracy tracking algorithms, however, remain scarce. In this paper we study the problem from a new perspective and present a novel parallel tracking and verifying (PTAV) framework, by taking advantage of the ubiquity of multi-thread techniques and borrowing ideas from the success of parallel tracking and mapping in visual SLAM. The proposed PTAV framework is typically composed of two components, a (base) tracker T and a verifier V, working in parallel on two separate threads. The tracker T aims to provide a super real-time tracking inference and is expected to perform well most of the time; by contrast, the verifier V validates the tracking results and corrects T when needed. The key innovation is that, V does not work on every frame but only upon the requests from T; on the other end, T may adjust the tracking according to the feedback from V. With such collaboration, PTAV enjoys both the high efficiency provided by T and the strong discriminative power by V. Meanwhile, to adapt V to object appearance changes over time, we maintain a dynamic target template pool for adaptive verification, resulting in further performance improvements. In our extensive experiments on popular benchmarks including OTB2015, TC128, UAV20L and VOT2016, PTAV achieves the best tracking accuracy among all real-time trackers, and in fact even outperforms many deep learning based algorithms. Moreover, as a general framework, PTAV is very flexible with great potentials for future improvement and generalization.
In this paper, we present a new method for detecting road users in an urban environment which leads to an improvement in multiple object tracking. Our method takes as an input a foreground image and improves the object detection and segmentation. This new image can be used as an input to trackers that use foreground blobs from background subtraction. The first step is to create foreground images for all the frames in an urban video. Then, starting from the original blobs of the foreground image, we merge the blobs that are close to one another and that have similar optical flow. The next step is extracting the edges of the different objects to detect multiple objects that might be very close (and be merged in the same blob) and to adjust the size of the original blobs. At the same time, we use the optical flow to detect occlusion of objects that are moving in opposite directions. Finally, we make a decision on which information we keep in order to construct a new foreground image with blobs that can be used for tracking. The system is validated on four videos of an urban traffic dataset. Our method improves the recall and precision metrics for the object detection task compared to the vanilla background subtraction method and improves the CLEAR MOT metrics in the tracking tasks for most videos.
Object tracking is one of the most challenging task and has secured significant attention of computer vision researchers in the past two decades. Recent deep learning based trackers have shown good performance on various tracking challenges. A tracking method should track objects in sequential frames accurately in challenges such as deformation, low resolution, occlusion, scale and light variations. Most trackers achieve good performance on specific challenges instead of all tracking problems, hence there is a lack of general purpose tracking algorithms that can perform well in all conditions. Moreover, performance of tracking techniques has not been evaluated in noisy environments. Visual object tracking has real world applications and there is good chance that noise may get added during image acquisition in surveillance cameras. We aim to study the robustness of two state of the art trackers in the presence of noise including Efficient Convolutional Operators (ECO) and Correlation Filter Network (CFNet). Our study demonstrates that the performance of these trackers degrades as the noise level increases, which demonstrate the need to design more robust tracking algorithms.
It is an important task to reliably detect and track multiple moving objects for video surveillance and monitoring. However, when occlusion occurs in nonlinear motion scenarios, many existing methods often fail to continuously track multiple moving objects of interest. In this paper we propose an effective approach for detection and tracking of multiple moving objects with occlusion. Moving targets are initially detected using a simple yet efficient block matching technique, providing rough location information for multiple object tracking. More accurate location information is then estimated for each moving object by a nonlinear tracking algorithm. Considering the ambiguity caused by the occlusion among multiple moving objects, we apply an unscented Kalman filtering (UKF) technique for reliable object detection and tracking. Different from conventional Kalman filtering (KF), which cannot achieve the optimal estimation in nonlinear tracking scenarios, UKF can be used to track both linear and nonlinear motions due to the unscented transform. Further, it estimates the velocity information for each object to assist to the object detection algorithm, effectively delineating multiple moving objects of occlusion. The experimental results demonstrate that the proposed method can correctly detect and track multiple moving objects with nonlinear motion patterns and occlusions.
Psychlab is a simulated psychology laboratory inside the first-person 3D game world of DeepMind Lab (Beattie et al. 2016). Psychlab enables implementations of classical laboratory psychological experiments so that they work with both human and artificial agents. Psychlab has a simple and flexible API that enables users to easily create their own tasks. As examples, we are releasing Psychlab implementations of several classical experimental paradigms including visual search, change detection, random dot motion discrimination, and multiple object tracking. We also contribute a study of the visual psychophysics of a specific state-of-the-art deep reinforcement learning agent: UNREAL (Jaderberg et al. 2016). This study leads to the surprising conclusion that UNREAL learns more quickly about larger target stimuli than it does about smaller stimuli. In turn, this insight motivates a specific improvement in the form of a simple model of foveal vision that turns out to significantly boost UNREAL's performance, both on Psychlab tasks, and on standard DeepMind Lab tasks. By open-sourcing Psychlab we hope to facilitate a range of future such studies that simultaneously advance deep reinforcement learning and improve its links with cognitive science.
Discrete correlation filter (DCF) based trackers have shown considerable success in visual object tracking. These trackers often make use of low to mid level features such as histogram of gradients (HoG) and mid-layer activations from convolution neural networks (CNNs). We argue that including semantically higher level information to the tracked features may provide further robustness to challenging cases such as viewpoint changes. Deep salient object detection is one example of such high level features, as it make use of semantic information to highlight the important regions in the given scene. In this work, we propose an improvement over DCF based trackers by combining saliency based and other features based filter responses. This combination is performed with an adaptive weight on the saliency based filter responses, which is automatically selected according to the temporal consistency of visual saliency. We show that our method consistently improves a baseline DCF based tracker especially in challenging cases and performs superior to the state-of-the-art. Our improved tracker operates at 9.3 fps, introducing a small computational burden over the baseline which operates at 11 fps.
We present a technique for simultaneous 3D reconstruction of static regions and rigidly moving objects in a scene. An RGB-D frame is represented as a collection of features, which are points and planes. We classify the features into static and dynamic regions and grow separate maps, static and object maps, for each of them. To robustly classify the features in each frame, we fuse multiple RANSAC-based registration results obtained by registering different groups of the features to different maps, including (1) all the features to the static map, (2) all the features to each object map, and (3) subsets of the features, each forming a segment, to each object map. This multi-group registration approach is designed to overcome the following challenges: scenes can be dominated by static regions, making object tracking more difficult; and moving object might have larger pose variation between frames compared to the static regions. We show qualitative results from indoor scenes with objects in various shapes. The technique enables on-the-fly object model generation to be used for robotic manipulation.
Visual object tracking is an important computer vision problem with numerous real-world applications including human-computer interaction, autonomous vehicles, robotics, motion-based recognition, video indexing, surveillance and security. In this paper, we aim to extensively review the latest trends and advances in the tracking algorithms and evaluate the robustness of trackers in the presence of noise. The first part of this work comprises a comprehensive survey of recently proposed tracking algorithms. We broadly categorize trackers into correlation filter based trackers and the others as non-correlation filter trackers. Each category is further classified into various types of trackers based on the architecture of the tracking mechanism. In the second part of this work, we experimentally evaluate tracking algorithms for robustness in the presence of additive white Gaussian noise. Multiple levels of additive noise are added to the Object Tracking Benchmark (OTB) 2015, and the precision and success rates of the tracking algorithms are evaluated. Some algorithms suffered more performance degradation than others, which brings to light a previously unexplored aspect of the tracking algorithms. The relative rank of the algorithms based on their performance on benchmark datasets may change in the presence of noise. Our study concludes that no single tracker is able to achieve the same efficiency in the presence of noise as under noise-free conditions; thus, there is a need to include a parameter for robustness to noise when evaluating newly proposed tracking algorithms.
Data association is a key step within the multi-object tracking pipeline that is notoriously challenging due to its combinatorial nature. A popular and general way to formulate data association is as the NP-hard multidimensional assignment problem (MDAP). Over the last few years, data-driven approaches to assignment have become increasingly prevalent as these techniques have started to mature. We focus this survey solely on learning algorithms for the assignment step of multi-object tracking, and we attempt to unify various methods by highlighting their connections to linear assignment as well as to the MDAP. First, we review probabilistic and end-to-end optimization approaches to data association, followed by methods that learn association affinities from data. We then compare the performance of the methods presented in this survey, and conclude by discussing future research directions.
In this paper, we propose a novel effective non-rigid object tracking framework based on the spatial-temporal consistent saliency detection. In contrast to most existing trackers that utilize a bounding box to specify the tracked target, the proposed framework can extract accurate regions of the target as tracking outputs. It achieves a better description of the non-rigid objects and reduces the background pollution for the tracking model. Furthermore, our model has several unique features. First, a tailored fully convolutional neural network (TFCN) is developed to model the local saliency prior for a given image region, which not only provides the pixel-wise outputs but also integrates the semantic information. Second, a novel multi-scale multi-region mechanism is proposed to generate local saliency maps that effectively consider visual perceptions with different spatial layouts and scale variations. Subsequently, local saliency maps are fused via a weighted entropy method, resulting in a final discriminative saliency map. Finally, we present a non-rigid object tracking algorithm based on the predicted saliency maps. By utilizing a spatial-temporal consistent saliency map (STCSM), we conduct target-background classification and use a simple fine-tuning scheme for online updating. Extensive experiments demonstrate that the proposed algorithm achieves competitive performance in both saliency detection and visual tracking, especially outperforming other related trackers on the non-rigid object tracking datasets.
Observing that Semantic features learned in an image classification task and Appearance features learned in a similarity matching task complement each other, we build a twofold Siamese network, named SA-Siam, for real-time object tracking. SA-Siam is composed of a semantic branch and an appearance branch. Each branch is a similarity-learning Siamese network. An important design choice in SA-Siam is to separately train the two branches to keep the heterogeneity of the two types of features. In addition, we propose a channel attention mechanism for the semantic branch. Channel-wise weights are computed according to the channel activations around the target position. While the inherited architecture from SiamFC \cite{SiamFC} allows our tracker to operate beyond real-time, the twofold design and the attention mechanism significantly improve the tracking performance. The proposed SA-Siam outperforms all other real-time trackers by a large margin on OTB-2013/50/100 benchmarks.