Visual tracking is one of the fundamental problems in computer vision. Recently, some deep-learning-based tracking algorithms have been illustrating record-breaking performances. However, due to the high complexity of neural networks, most deep trackers suffer from low tracking speed and are, thus, impractical in many real-world applications. Some recently proposed deep trackers with smaller network structure achieve high efficiency while at the cost of significant decrease in precision. In this paper, we propose to transfer the deep feature, which is learned originally for image classification to the visual tracking domain. The domain adaptation is achieved via some “grafted” auxiliary networks, which are trained by regressing the object location in tracking frames. This adaptation improves the tracking performance significantly both on accuracy and efficiency. The yielded deep tracker is real time and also illustrates the state-of-the-art accuracies in the experiment involving two well-adopted benchmarks with more than 100 test videos. Furthermore, the adaptation is also naturally used for introducing the objectness concept into visual tracking. This removes a long-standing target ambiguity in visual tracking tasks, and we illustrate the empirical superiority of the more well-defined task.

We are now living in an era of visual information where data is unceasingly generated and pushed into consumption at astounding rates. A remarkable portion of this sensory input comes in the form of videos streaming from large-scale surveillance infrastructures as well as consumer-grade monitoring systems. The sheer amount of ground-based, aerial and mobile video surveillance data demands fittingly competent, accurate, effective techniques to extract useful cues and provide assistance for detection, prevention, and intervention tasks in traffic, safety, security, defense, forensic, health, biology, ethology, and retail space management applications.

Visual tracking is a fundamental problem in computer vision. However, due to the (sometimes) ambiguous target information given at the first frame, it has also been criticized as less well-posed compared with other tasks with clearly-defined targets, such as object detection and semantic segmentation. In this paper, we try to evaluate the importance of object category in visual tracking by tracking objects with known object types. The proposed algorithm, termed Deep-Track with Objectness (DTO), naturally combines the state-of-the-art deep-learning-based detectors and trackers, which essentially share a large part of the network. In DTO, a deep tracker, which is scale-fixed and sensitive to small translations tracks the object in a relative short lifespan. A deep detector, which is scale-changeable and robust to pose or illumination changes guides the deep tracker in a longer lifespan. As the deep tracker and detector share the main part of their networks, no much extra computation is imposed while the performance gain is significant. We test the proposed algorithm on two well-accepted benchmarks and on both of them, the proposed method increases the tracking accuracies remarkably compared with state-of-the-art visual trackers.

Visual tracking is a fundamental problem in computer vision. Recently, some deep-learning-based tracking algorithms have been achieving record-breaking performances. However, due to the high complexity of deep learning, most deep trackers suffer from low tracking speed, and thus are impractical in many real-world applications. Some new deep trackers with smaller network structure achieve high efficiency while at the cost of significant decrease on precision. In this paper, we propose to transfer the feature for image classification to the visual tracking domain via convolutional channel reductions. The channel reduction could be simply viewed as an additional convolutional layer with the specific task. It not only extracts useful information for object tracking but also significantly increases the tracking speed. To better accommodate the useful feature of the target in different scales, the adaptation filters are designed with different sizes. The yielded visual tracker is real-time and also illustrates the state-of-the-art accuracies in the experiment involving two well-adopted benchmarks with more than 100 test videos.

Inferring the aesthetic quality of images is a challenging computer vision task due to its subjective and conceptual nature. Most image aesthetics evaluation approaches focused on designing handcrafted features, and only a few adopted learning of relevant and imperative characteristics in a data-driven manner. In this paper, we propose to attune Convolutional Neural Networks (CNNs) for image aesthetics. Unlike previous deep learning based techniques, we employ pretrained models, namely AlexNet [12] and the 16-layer VGGNet [20], and calibrate them to estimate visual aesthetic quality. This enables exploiting automatically the inherent information from much larger scale and more diversified image datasets. We tested our methods on AVA and CUHKPQ image aesthetics datasets on two different training-testing partitions, and compared the performance using both local and contextual information. Experimental results suggest that our strategy is robust, effective and superior to the state-of-the-art approaches.

Deep neural networks, albeit their great success on feature learning in various computer vision tasks, are usually considered as impractical for online visual tracking, because they require very long training time and a large number of training samples. In this paper, we present an efficient and very robust tracking algorithm using a single convolutional neural network (CNN) for learning effective feature representations of the target object in a purely online manner. Our contributions are multifold. First, we introduce a novel truncated structural loss function that maintains as many training samples as possible and reduces the risk of tracking error accumulation. Second, we enhance the ordinary stochastic gradient descent approach in CNN training with a robust sample selection mechanism. The sampling mechanism randomly generates positive and negative samples from different temporal distributions, which are generated by taking the temporal relations and label noise into account. Finally, a lazy yet effective updating scheme is designed for CNN training. Equipped with this novel updating algorithm, the CNN model is robust to some long-existing difficulties in visual tracking, such as occlusion or incorrect detections, without loss of the effective adaption for significant appearance changes. In the experiment, our CNN tracker outperforms all compared state-of-the-art methods on two recently proposed benchmarks, which in total involve over 60 video sequences. The remarkable performance improvement over the existing trackers illustrates the superiority of the feature representations, which are learned purely online via the proposed deep learning framework.

Retrieving 3D models from 2D human sketches has received considerable attention in the areas of graphics, image retrieval, and computer vision. Almost always in state of the art approaches a large amount of “best views” are computed for 3D models, with the hope that the query sketch matches one of these 2D projections of 3D models using predefined features. We argue that this two stage approach (view selection - matching) is pragmatic but also problematic because the “best views” are subjective and ambiguous, which makes the matching inputs obscure. This imprecise nature of matching further makes it challenging to choose features manually. Instead of relying on the elusive concept of “best views” and the hand-crafted features, we propose to define our views using a minimalism approach and learn features for both sketches and views. Specifically, we drastically reduce the number of views to only two predefined directions for the whole dataset. Then, we learn two Siamese Convolutional Neural Networks (CNNs), one for the views and one for the sketches. The loss function is defined on the within-domain as well as the cross-domain similarities. Our experiments on three benchmark datasets demonstrate that our method is significantly better than state of the art approaches, and outperforms them in all conventional metrics.

In this work we describe a compact multi-task Convolutional Neural Network (CNN) for simultaneously estimating image quality and identifying distortions. CNNs are natural choices for multi-task problems because learned convolutional features may be shared by different high level tasks. However, we empirically argue that simply appending additional tasks based on the state of the art structure (e.g., [1]) does not lead to optimal solutions. We design a compact structure with nearly 90% fewer parameters compared to [1], and demonstrate its learning power.

Matching hand drawn sketches is an attractive topic in image understanding and potentially has many applications. Previous sketch matching algorithms often rely on extracted feature points and their correspondence. However, the nature of hand drawn sketches, such as lack of constraints and having significantly large variations, makes the matching task extremely challenging. In this paper, we propose a metric learning method to match hand drawn sketches without explicitly localizing the feature points. We train a Siamese Convolutional Neural Network (CNN) with pure convolutional layers to represent the sketch features. This allows us to benefit from the rich representative power of CNN, as well as to preserve the spatial information of features. We evaluated the sketch retrieval performance of our model on a large dataset. Experiment results showed the effectiveness of our model.

In this paper, higher-order correlation clustering (HOCC) is used for text line detection in natural images. We treat text line detection as a graph partitioning problem, where each vertex is represented by a Maximally Stable Extremal Region (MSER). First, weak hypothesises are proposed by coarsely grouping MSERs based on their spatial alignment and appearance consistency. Then, higher-order correlation clustering (HOCC) is used to partition the MSERs into text line candidates, using the hypotheses as soft constraints to enforce long range interactions. We further propose a regularization method to solve the Semidefinite Programming problem in the inference. Finally we use a simple texton-based texture classifier to filter out the non-text areas. This framework allows us to naturally handle multiple orientations, languages and fonts. Experiments show that our approach achieves competitive performance compared to the state of the art.

In this work we describe a Convolutional Neural Network (CNN) to accurately predict image quality without a reference image. Taking image patches as input, the CNN works in the spatial domain without using hand-crafted features that are employed by most previous methods. The network consists of one convolutional layer with max and min pooling, two fully connected layers and an output node. Within the network structure, feature learning and regression are integrated into one optimization process, which leads to a more effective model for estimating image quality. This approach achieves state of the art performance on the LIVE dataset and shows excellent generalization ability in cross dataset experiments. Further experiments on images with local distortions demonstrate the local quality estimation ability of our CNN, which is rarely reported in previous literature.

This paper proposes a deep learning approach for document image quality assessment. Given a noise corrupted document image, we estimate its quality score as a prediction of OCR accuracy. First the document image is divided into patches and non-informative patches are sifted out using Otsu's binarization technique. Second, quality scores are obtained for all selected patches using a Convolutional Neural Network (CNN), and the patch scores are averaged over the image to obtain the document score. The proposed CNN contains two layers of convolution, location blind max-min pooling, and Rectified Linear Units in the fully connected layers. Experiments on two document quality datasets show our method achieved the state of the art performance.

This paper presents a Convolutional Neural Network (CNN) for document image classification. In particular, document image classes are defined by the structural similarity. Previous approaches rely on hand-crafted features for capturing structural information. In contrast, we propose to learn features from raw image pixels using CNN. The use of CNN is motivated by the the hierarchical nature of document layout. Equipped with rectified linear units and trained with dropout, our CNN performs well even when document layouts present large inner-class variations. Experiments on public challenging datasets demonstrate the effectiveness of the proposed approach.

Simple tree models for articulated objects prevails in the last decade. However, it is also believed that these simple tree models are not capable of capturing large variations in many scenarios, such as human pose estimation. This paper attempts to address three questions: 1) are simple tree models sufficient? more specifically, 2) how to use tree models effectively in human pose estimation? and 3) how shall we use combined parts together with single parts efficiently? Assuming we have a set of single parts and combined parts, and the goal is to estimate a joint distribution of their locations. We surprisingly find that no latent variables are introduced in the Leeds Sport Dataset (LSP) during learning latent trees for deformable model, which aims at approximating the joint distributions of body part locations using minimal tree structure. This suggests one can straightforwardly use a mixed representation of single and combined parts to approximate their joint distribution in a simple tree model. As such, one only needs to build Visual Categories of the combined parts, and then perform inference on the learned latent tree. Our method outperformed the state of the art on the LSP, both in the scenarios when the training images are from the same dataset and from the PARSE dataset. Experiments on animal images from the VOC challenge further support our findings.

Existing dual image deblurring methods usually model blurred image pairs being taken from exactly the same viewpoint and restore a single clear image. This imposes a strong assumption that the latent clear images of both images must be completely identical. In contrast to this restricted scenario, we assume that the restored pair are different, but can be approximated by image warping due to small viewpoint change. This allows us to deblur each image individually, but still being able to make use of the matched areas in image pairs. Our deblurring algorithm iteratively performs a two-directional dual image deblurring, which uses the Split Bregman method, and matches the latent clear image pairs by a homography. Experiments show that the proposed algorithm automatically recovers clear images from blurred image pairs in the same scene. Statistics suggest that the method is robust to viewpoint change and different noise levels.

Text line segmentation in freestyle handwritten documents remains an open document analysis problem. Curvilinear text lines and small gaps between neighboring text lines present a challenge to algorithms developed for machine printed or hand-printed documents. In this paper, we propose a novel approach based on density estimation and a state-of-the-art image segmentation technique, the level set method. From an input document image, we estimate a probability map, where each element represents the probability that the underlying pixel belongs to a text line. The level set method is then exploited to determine the boundary of neighboring text lines by evolving an initial estimate. Unlike connected component based methods ( [1], [2] for example), the proposed algorithm does not use any script-specific knowledge. Extensive quantitative experiments on freestyle handwritten documents with diverse scripts, such as Arabic, Chinese, Korean, and Hindi, demonstrate that our algorithm consistently outperforms previous methods [1]-[3]. Further experiments show the proposed algorithm is robust to scale change, rotation, and noise.