Nowadays, it is very convenient to capture photos by a smart phone. As using, the phone way share what users experienced anytime and anywhere through social networks, possible that we multiple make sure content well photographed. In this paper, an effective scalable mobile image retrieval approach proposed exploring contextual salient information for input query image. Our goal explore high-level semantic of finding saliency from relevant rather than solely using Thus, first determines...

Fine-grained food recognition is the detailed classification that provides more specialized and professional attribute information of food. It basic work to realize healthy diet recommendations cooking instructions, nutrition intake management, cafeteria self-checkout system. Chinese lacks structured information, ingredients composition an important consideration. The current approaches mostly focus on global dish appearance without any analysis ingredient fully considering attention...

Automated cell detection and localization from microscopy images are significant tasks in biomedical research clinical practice. In this paper, we design a new algorithm that combines deep convolutional neural network (CNN) compressed sensing (CS) or sparse coding (SC) for end-to-end training. We also derive, the first time, backpropagation rule, which is applicable to train any implements code recovery layer. The key innovation behind our task structured as point object computer vision,...

The number of mitotic cells present in histopathological slides is an important predictor tumor proliferation the diagnosis breast cancer. However, current approaches can hardly perform precise pixel-level prediction for mitosis datasets with only weak labels (i.e., provide centroid location cells), and take no account large domain gap across from different pathology laboratories. In this work, we propose a Domain adaptive Box-supervised Instance segmentation Network (DBIN) to address above...

Landmark summarization with diverse viewpoints is very important in landmark retrieval, as it can create a comprehensive description of for users. In this paper we present an approach summarizing collection images from viewpoints. First, group content overlap by viewpoint album (VA) generation. Second, model the relative each image within VA based on spatial layout distinctive descriptors landmark. Third, express 4-D vector, including horizontal, vertical, scale, and rotation. Finally,...

The ability to automatically detect certain types of cells or cellular subunits in microscopy images is significant interest a wide range biomedical research and clinical practices. Cell detection methods have evolved from employing hand-crafted features deep learning-based techniques. essential idea these that their cell classifiers detectors are trained the pixel space, where locations target labeled. In this paper, we seek different route propose convolutional neural network (CNN)-based...

Crowd scene analysis receives growing attention due to its wide applications. Grasping the accurate crowd location is important for identifying high-risk regions. In this article, we propose a Compressed Sensing based Output Encoding (CSOE) scheme, which casts detecting pixel coordinates of small objects into task signal regression in encoding space. To prevent gradient vanishing, derive our own sparse reconstruction backpropagation rule that adaptive distinct implementations and makes whole...

In this paper, we describe an approach for visually summarizing a landmark by recommending images with diverse viewpoints (e.g. front-side viewpoint, bottom-top close-distant etc). Our models image's viewpoint using 4-D vector, which describes in horizontal, vertical, scale and orientation aspects. To construct the vector image, select Identical Semantic Points (ISPs) from hundreds to thousands SIFT points of image captures some major unique parts landmark. Then four dimensional is utilized...

Representative images generation offers a comprehensive knowledge for landmark and is hot research area recent years. This paper presents representative system by discovering high frequency shooting locations from geo-tagged community-contributed photos. We discover that the views (e.g. far near, front, back side) of photos taken in same location are usually similar but different locations. Our realized three steps: 1) Landmark dataset filtered social media combination tags geo-tags. 2) High...

With the increasing popularity of intelligent surveillance systems, abnormal behavior detection human beings based on computer vision is attracting more attention. It aims to classify and locate behaviors coordinates beings, respectively, a fundamental technology for security. Existing approaches mainly focus exploring features through object detectors. However, in office scenarios, almost all are closely associated with fine-grained feature around nose, wrist, elbow, other joint points...

In this paper, we propose a novel image retrieval scheme, where multi relevant images are input as queries to improve the performance. We exploit sufficient information provided by query reduce distractor features, quantization loss and learn visual synonyms. During learning synonyms, consisting of synonyms detection expansion, some identical unique details semantically important captured. represent using set each which comprises several word paths, quantizing descriptor from root leaf...

Silicone‐based dielectric elastomers are promising electroactive polymers (DEAPs) applicable to various actuator applications. However, the lack of information concerning their long‐term performance still limits industrial use. Here, time‐dependent behavior silicon‐based DEAPs under electromechanical cycling is investigated. A series thin silicone films prepared with different stoichiometric imbalances coated compliant silver nanowire electrodes and then electromechanically cycled...

Surgical tool localization is the foundation to a series of advanced surgical functions <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.</i> image guided navigation. For precise scenarios like localization, sophisticated tools and sensitive tissues can be quite close. This requires higher accuracy than general object localization. And it also meaningful know orientation tools. To achieve these, this paper proposes Compressive Sensing...

Recently, the emerging concept of "unmanned retail" has drawn more and attention, unmanned retail based on intelligent vending machines (UVMs) scene great market demand. However, existing product recognition methods for UVMs cannot adapt to large-scale categories have insufficient accuracy. In this article, we propose a method UVMs. It can be divided into two parts: 1) first, explore similarities differences between products through manifold learning, then build hierarchical multigranularity...

Synthetic aperture radar (SAR) is widely used in terrain classification, object detection, and other fields. Compared with anchor-based detectors, anchor-free detectors remove the anchor mechanism implement detection box encoding a more elegant form. However, are limited by complex scenes caused geometric transformations, such as overlaying, shadow, vertex displacement during SAR imaging. And scattered power distribution of noise similar to edge object, making it difficult for detector...

The ability to detect certain types of cells in a microscopy image is important for wide range clinical applications. Cells often present huge variations density and appearance, occupy only small portion an image. Consequently, general object detection methods computer vision do not meet accuracy requirements: false / missed detections prevail. In this paper, we apply convolutional neural network (CNN) regress fixed length vector from Then, L <sub...