Search

We study the problem of 3D object detection from RGB-D images so as to achieve localization (i.e., producing a bounding box around the object) and classification (i.e., determining the object category) simultaneously. Its challenges arise from high intra-class variability, illumination change, background clutter and occlusion. To solve this problem, we propose a novel solution that integrates the 2D information (RGB images), the 3D information (RGB-D images) and the object/scene context information together, and call it the Context-Assisted 3D (C3D) method. In the proposed C3D method, we first use a convolutional neural network (CNN) to jointly detect a 3D object in a scene and its scene category. Then, we improve the detection result furthermore with a Conditional Random Field (CRF) model that incorporates the object potential, the scene potential, the scene/object context, the object/object context, and the room geometry. Extensive experiments are conducted to demonstrate that the proposed C3D method achieves the state-of-the-art performance for 3D object detection against the SUN RGB-D benchmark dataset. [ABSTRACT FROM AUTHOR]

In this work, we propose a technique that utilizes a fully convolutional network (FCN) to localize image splicing attacks. We first evaluated a single-task FCN (SFCN) trained only on the surface label. Although the SFCN is shown to provide superior performance over existing methods, it still provides a coarse localization output in certain cases. Therefore, we propose the use of a multi-task FCN (MFCN) that utilizes two output branches for multi-task learning. One branch is used to learn the surface label, while the other branch is used to learn the edge or boundary of the spliced region. We trained the networks using the CASIA v2.0 dataset, and tested the trained models on the CASIA v1.0, Columbia Uncompressed, Carvalho, and the DARPA/NIST Nimble Challenge 2016 SCI datasets. Experiments show that the SFCN and MFCN outperform existing splicing localization algorithms, and that the MFCN can achieve finer localization than the SFCN. [ABSTRACT FROM AUTHOR]

Inflammatory pleuritis often causes pleural effusions, which are drained through lymphatic vessels (lymphatics) in the parietal pleura. The distribution of button- and zipper-like endothelial junctions can identify the subtypes of lymphatics, the initial, pre-collecting, and collecting lymphatics. Vascular endothelial growth factor receptor (VEGFR)-3 and its ligands VEGF-C/D are crucial lymphangiogenic factors. Currently, in the pleura covering the chest walls, the anatomy of the lymphatics and connecting networks of blood vessels are incompletely understood. Moreover, their pathological and functional plasticity under inflammation and the effects of VEGFR inhibition are unclear. This study aimed to learn the above-unanswered questions and immunostained mouse chest walls as whole-mount specimens. Confocal microscopic images and their 3-dimensional reconstruction analyzed the vasculatures. Repeated intra-pleural cavity lipopolysaccharide challenge induced pleuritis, which was also treated with VEGFR inhibition. Levels of vascular-related factors were evaluated by quantitative real-time polymerase chain reaction. We observed the initial lymphatics in the intercostals, collecting lymphatics under the ribs, and pre-collecting lymphatics connecting both. Arteries branched into capillaries and gathered into veins from the cranial to the caudal side. Lymphatics and blood vessels were in different layers with an adjacent distribution of the lymphatic layer to the pleural cavity. Inflammatory pleuritis elevated expression levels of VEGF-C/D and angiopoietin-2, induced lymphangiogenesis and blood vessel remodeling, and disorganized the lymphatic structures and subtypes. The disorganized lymphatics showed large sheet-like structures with many branches and holes inside. Such lymphatics were abundant in zipper-like endothelial junctions with some button-like junctions. The blood vessels were tortuous and had various diameters and complex networks. Stratified layers of lymphatics and blood vessels were disorganized, with impaired drainage function. VEGFR inhibition partially maintained their structures and drainage function. These findings demonstrate anatomy and pathological changes of the vasculatures in the parietal pleura and their potential as a novel therapeutic target. • Anatomy of the lymphatics is unclear in the pleura covering the chest wall. • We found collecting lymphatics under the ribs and the initials in the intercostals. • Inflammatory pleuritis induced lymphangiogenesis and blood vessel remodeling. • Pleuritis damaged the lymphatic structures and impaired their drainage function. • VEGFR inhibition maintained the lymphatic structures and drainage function. [ABSTRACT FROM AUTHOR]

An approach that extracts global attributes from outdoor images to facilitate geometric layout labeling is investigated in this work. The proposed Global-attributes Assisted Labeling (GAL) system exploits both local features and global attributes. First, by following a classical method, we use local features to provide initial labels for all super-pixels. Then, we develop a set of techniques to extract global attributes from 2D outdoor images. They include sky lines, ground lines, vanishing lines, etc. Finally, we propose the GAL system that integrates global attributes in the conditional random field (CRF) framework to improve initial labels so as to offer a more robust labeling result. The performance of the proposed GAL system is demonstrated and benchmarked with several state-of-the-art algorithms against a popular outdoor scene layout dataset. [ABSTRACT FROM AUTHOR]