Filters








5,414 Hits in 7.4 sec

Pixel-Level Hand Detection with Shape-Aware Structured Forests [chapter]

Xiaolong Zhu, Xuhui Jia, Kwan-Yee K. Wong
2015 Lecture Notes in Computer Science  
In this paper, we propose a novel method for efficient pixel-level hand detection.  ...  Unlike previous method which assigns a binary label to every pixel independently, our method estimates a probability shape mask for a pixel using structured forests.  ...  Shape-aware Structured Forests In this section, we first briefly review the pixel-level hand detection using random forests.  ... 
doi:10.1007/978-3-319-16817-3_5 fatcat:kl23kn6qunbjfd3cqplltgghby

A two-stage detector for hand detection in ego-centric videos

Xiaolong Zhu, Wei Liu, Xuhui Jia, Kwan-Yee K. Wong
2016 2016 IEEE Winter Conference on Applications of Computer Vision (WACV)  
In the first stage, hand bounding box proposals are generated from a pixel-level hand probability map.  ...  Through experiments, we demonstrate that our method is efficient and robust to detect hands with their shape and landmark information, and our system can also be flexibly combined with other detection  ...  Previous pixel-level hand detection approaches [29, 12] enable us to obtain a probability map of hand pixels efficiently under a random forest framework.  ... 
doi:10.1109/wacv.2016.7477665 dblp:conf/wacv/ZhuLJW16 fatcat:5uynppgb3jhb7i6ux5soqqpkvm

Joint Estimation of 3D Hand Position and Gestures from Monocular Video for Mobile Interaction

Jie Song, Fabrizio Pece, Gábor Sörös, Marion Koelle, Otmar Hilliges
2015 Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems - CHI '15  
We also classify distinct hand gestures, allowing for a variety of 3D interactions.  ...  We demonstrate our technique with three mobile interaction scenarios and evaluate the method quantitatively and qualitatively.  ...  We extend [13] to jointly detect hand shapes and to regress the average hand depth. Together with the hand's centroid this gives full 3D hand position.  ... 
doi:10.1145/2702123.2702601 dblp:conf/chi/SongPSKH15 fatcat:2ytudpomezcnxpokxhmbdvzmn4

Unified Face Analysis by Iterative Multi-output Random Forests

Xiaowei Zhao, Tae-Kyun Kim, Wenhan Luo
2014 2014 IEEE Conference on Computer Vision and Pattern Recognition  
On one hand, the estimated pose and expression provide strong shape prior to constrain the variation of landmark positions.  ...  Specifically, a hierarchical face analysis forest is learned to perform classification of pose and expression at the top level, while performing landmark positions regression at the bottom level.  ...  Acknowledgments This work is supported by EPSRC grant (EP/J012106/1) 3D intrinsic shape recognition.  ... 
doi:10.1109/cvpr.2014.228 dblp:conf/cvpr/ZhaoKL14 fatcat:ictgpdzzf5dkjbqxl6vt57syby

Superpixel-based structure classification for laparoscopic surgery

Sebastian Bodenstedt, Jochen Görtler, Martin Wagner, Hannes Kenngott, Beat Peter Müller-Stich, Rüdiger Dillmann, Stefanie Speidel, Robert J. Webster, Ziv R. Yaniv
2016 Medical Imaging 2016: Image-Guided Procedures, Robotic Interventions, and Modeling  
Using a Random Forest classifier, we determine the class label of each superpixel.  ...  First, a given laparoscopic image is divided into groups of connected pixels, so-called superpixels, using the SEEDS algorithm.  ...  the SEEDS algorithm outlined in, 6 due to it being able to compute s(i) in real-time (30Hz with an image resolution of 640x480 pixels), while taking homogeneity and shape of potential superpixel candidate  ... 
doi:10.1117/12.2216750 dblp:conf/miigp/BodenstedtGWKMD16 fatcat:culmsyjb4rafnjqtdr25kvk6mm

Detection of Microaneurysms and Hemorrhages in Fundus Image for Glaucoma Diagnosis

R Shanthi, S Prabakaran
2018 International Journal of Engineering & Technology  
In future, this methodology can be applied to larger clinical dataset in order to detect the disease at the maximum level.  ...  The process of screening is evaluated on publicly available database Diaretdb1 with the resolution of 1152 x 1500 pixels for both healthy and abnormal images.  ...  Automation can be attained in two levels: first level is segmenting the blood vessels to detect the fundus abnormalities and the second level is classifying the severity level.  ... 
doi:10.14419/ijet.v7i2.24.12092 fatcat:i2oce5chm5eovay3djzz4vkhyq

Analysis of the hands in egocentric vision: A survey [article]

Andrea Bandini, José Zariffa
2020 arXiv   pre-print
Moreover, a list of the most prominent datasets with hand-based annotations is provided.  ...  This intrinsic advantage enables the study of the hands from multiple perspectives: localizing hands and their parts within the images; understanding what actions and activities the hands are involved  ...  [39] , [40] learned the segmentation mask within the image patch, by using a random forest framework (i.e., shape-aware structured forest).  ... 
arXiv:1912.10867v2 fatcat:rd2cu3jbkvbvnii5s7fh5c2ncq

Pushing the Limits of Deep CNNs for Pedestrian Detection [article]

Qichang Hu, Peng Wang, Chunhua Shen, Anton van den Hengel, Fatih Porikli
2016 arXiv   pre-print
We also show that pixel labelling may be simply combined with a detector to boost the detection performance.  ...  A breakthrough was made very recently by using sophisticated deep CNN models, with a number of hand-crafted features, or explicit occlusion handling mechanism.  ...  The average precision summaries the shape of the precision-recall curve, and is defined as the mean precision at a set of evenly spaced recall levels.  ... 
arXiv:1603.04525v2 fatcat:r2b4meixcvd25bvrvw62k7stbe

Left/right hand segmentation in egocentric videos

Alejandro Betancourt, Pietro Morerio, Emilia Barakova, Lucio Marcenaro, Matthias Rauterberg, Carlo Regazzoni
2017 Computer Vision and Image Understanding  
cooperative entities, ii) close hand interactions may lead to hand-to-hand occlusions and, as a consequence, create a single hand-like segment.  ...  Hand-to-hand occlusions are addressed by exploiting temporal superpixels.  ...  to the shape-aware classifier proposed in [27] .  ... 
doi:10.1016/j.cviu.2016.09.005 fatcat:hvp4adgxrnh3ve5agj4cy6vyii

SEMANTIC SEGMENTATION OF AERIAL IMAGES IN URBAN AREAS WITH CLASS-SPECIFIC HIGHER-ORDER CLIQUES

J. A. Montoya-Zegarra, J. D. Wegner, L. Ladický, K. Schindler
2015 ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences  
To specifically model the structure of roads and buildings, we add high-level shape representations for both classes by sampling large sets of putative object candidates.  ...  To obtain the final pixel-wise labeling, we use a CRF with higher-order potentials that balances the data term with the object candidates. We achieve overall labeling accuracies of > 80%.  ...  On the one hand, the path cliques close road network gaps, while on the other hand α-shape cliques repair building shapes.  ... 
doi:10.5194/isprsannals-ii-3-w4-127-2015 fatcat:3iv2i7kr7rddxe7wryqm5ljgeu

Semantic-Aware Image Smoothing [article]

Weihao Li, Omid Hosseini Jafari, Carsten Rother
2017 International Symposium on Vision, Modeling, and Visualization  
In this paper, we present an interesting observation, i.e. high-level semantic image labeling information can provide a meaningful structure prior naturally.  ...  On the other hand, semantic image labeling has achieved significant progress recently and has been widely used in many computer vision tasks.  ...  Structured edge detection [DZ13] is a popular edge detection method based on random forests, which can detect salient edges.  ... 
doi:10.2312/vmv.20171271 dblp:conf/vmv/LiJR17 fatcat:4ttcrlb3abcszhwh7esqlpgyuu

Category-level 6D Object Pose Recovery in Depth Images [article]

Caner Sahin, Tae-Kyun Kim
2018 arXiv   pre-print
Our architecture particularly adapts the distribution shifts arising from shape discrepancies, and naturally removes the variations of texture, illumination, pose, etc., so we call it as "Intrinsic Structure  ...  We engineer ISA based on the followings: i) "Semantically Selected Centers (SSC)" are proposed in order to define the "6D pose" at the level of categories. ii) 3D skeleton structures, which we derive as  ...  For the reader's convenience, we only review 6D case for instance-level object detection and pose estimation, and keep category-level detection broader.  ... 
arXiv:1808.00255v1 fatcat:hlmdfrymdbaprnbxd4pdvw2rwy

Detection of Hedges in a Rural Landscape Using a Local Orientation Feature: From Linear Opening to Path Opening

Mathieu Fauvel, Benoit Arbelot, Jón Atli Benediktsson, David Sheeren, Jocelyn Chanussot
2013 IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing  
From the membership map, the local orientation of each pixel is computed using directional filters.  ...  The detection of hedges is addressed in this paper. A hierarchical detection scheme in two steps is proposed.  ...  The second image illustrate the problem of forests with peculiar shapes, and close to hedges.  ... 
doi:10.1109/jstars.2012.2227464 fatcat:ob3jlry7ynewtg4xntvnjsvpuy

Towards Sustainable Land Uses within the Elbe River Biosphere Reserve in Lower Saxony, Germany by Means of TerraSAR-X Images

Dalia Farghaly, Emad Elba, Brigitte Urban
2016 Journal of Geoscience and Environment Protection  
Those areas are characterized by several types of land use, with agricultural land use being predominant.  ...  During the segmentation process, pixels with similar feature values (brightness, texture, color, shape) are grouped into small objects.  ...  Finally, the classification results can be exported to shape files and/or raster images [54] . The Decision-Tree method builds classification in the form of a tree structure.  ... 
doi:10.4236/gep.2016.43009 fatcat:nck7p2egkzdhzeg37krsi2nlym

Detecting dominant landscape objects through multiple scales: An integration of object-specific methods and watershed segmentation

Ola Hall, Geoffrey J. Hay, André Bouchard, Danielle J. Marceau
2004 Landscape Ecology  
To detect significant features occurring at specific levels of organization in a landscape, two steps are required. First, a multiscale dataset must be generated from which these features can emerge.  ...  This study was conducted on an agro-forested region in southwest Quebec, Canada, using IKONOS satellite data.  ...  Thus, the user must be aware that the types of image-objects that will be defined are based on the relationship between pixel grain, the size/shape of the analyzing kernel͑s͒, and the size of the image-objects  ... 
doi:10.1023/b:land.0000018371.43447.1f fatcat:l7qwj4bmu5gnnasyrxkz5ojbjm
« Previous Showing results 1 — 15 out of 5,414 results