WO2013144418A1 - Segmentation d'image - Google Patents

Segmentation d'image Download PDF

Info

Publication number
WO2013144418A1
WO2013144418A1 PCT/FI2012/050313 FI2012050313W WO2013144418A1 WO 2013144418 A1 WO2013144418 A1 WO 2013144418A1 FI 2012050313 W FI2012050313 W FI 2012050313W WO 2013144418 A1 WO2013144418 A1 WO 2013144418A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
foreground object
user command
location
segmentation process
Prior art date
Application number
PCT/FI2012/050313
Other languages
English (en)
Inventor
Kemal Ugur
Aydin ALATAN
Erhan GUNDOGDU
Ozan SENER
Original Assignee
Nokia Corporation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nokia Corporation filed Critical Nokia Corporation
Priority to PCT/FI2012/050313 priority Critical patent/WO2013144418A1/fr
Publication of WO2013144418A1 publication Critical patent/WO2013144418A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/143—Segmentation; Edge detection involving probabilistic approaches, e.g. Markov random field [MRF] modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/162—Segmentation; Edge detection involving graph-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
    • G06V10/422—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation for representing the structure of the pattern or shape of an object therefor
    • G06V10/426—Graphical representations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20072—Graph-based image processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20076—Probabilistic image processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20092—Interactive image processing based on input by user
    • G06T2207/20096—Interactive definition of curve of interest
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20092—Interactive image processing based on input by user
    • G06T2207/20104—Interactive definition of region of interest [ROI]

Definitions

  • the present invention relates to image processing, and more particularly to a process of image segmentation Background of the invention
  • 3D content Stereoscopy and displaying three-dimensional (3D) content has been a major research area for decades. Consequently, various stereoscopic displays of different sizes has been proposed and implemented. Big sizes of 3D displays have already gained popularity and mobile displays are expected to be popular soon. Despite of mobile 3D displays being ready for market, there is very little 3D content available for these displays. While waiting for more 3D content to be available in future, conversion of 2D content to 3D content may partially solve the problem in the intermediate period.
  • a method according to the invention is based on the idea of:
  • the method further comprises:
  • the method further comprises: segmenting, prior to step b, the image into over-segments preserving a graph structure and spatial neighbourhood of pixel data of the image; and carrying out the segmentation process on an over-segment level.
  • the method further comprises: providing a grayscale version of the image on the display; and displaying the foreground object in color during the segmentation process.
  • the segmentation process is carried out within a spatial neighbourhood of the foreground object.
  • the method further comprises: displaying the mask image as a touch of a brush being centered at the location of the user command and having a predefined radius.
  • the method further comprises: in response to a location of the user command extending beyond boundaries of the foreground object, marking the location as an exit point; in response to a location of the user command returning within the boundaries of the foreground object, marking the location as an entrance point; and amending the path of the mask image to be between the exit point and the entrance point along the boundary of the foreground object.
  • the method further comprises: adjusting the radius of the brush according to local image characteristics around the location of the user command.
  • the method further comprises: in response to detecting that no further user commands are provided on the foreground object, completing the segmentation process of the image on the basis of the information available in the updated classification algorithm.
  • an apparatus comprising at least one processor, memory including computer program code, the memory and the computer program code configured to, with the at least one processor, cause the apparatus to at least: a) provide an image on a display;
  • a computer readable storage medium stored with code thereon for use by an apparatus, which when executed by a processor, causes the apparatus to perform:
  • FIG. 1 a and 1 b show a system and devices suitable to be used in a image segmentation according to an embodiment
  • Fig. 2 shows a flow chart of a image segmentation method according to an embodiment of the invention
  • Figs. 3a, 3b and 3c illustrate an example of user interaction according to an embodiment of the invention
  • Fig. 4 illustrate an example of user interaction according to another embodiment of the invention
  • Fig. 5 illustrate an example of user interaction according to a further embodiment of the invention
  • Figs. 6a and 6b illustrate the propagation of the user interaction from a first frame of a video sequence to a second frame.
  • Figs. 1 a and 1 b show a system and devices suitable to be used in an image segmentation according to an embodiment.
  • the different devices may be connected via a fixed network 21 0 such as the Internet or a local area network; or a mobile communication network 220 such as the Global System for Mobile communications (GSM) network, 3rd Generation (3G) network, 3.5th Generation (3.5G) network, 4th Generation (4G) network, Wireless Local Area Network (WLAN), Bluetooth ® , or other contemporary and future networks.
  • GSM Global System for Mobile communications
  • 3G 3rd Generation
  • 3.5G 3.5th Generation
  • 4G 4th Generation
  • WLAN Wireless Local Area Network
  • Bluetooth ® Wireless Local Area Network
  • the networks comprise network elements such as routers and switches to handle data, and communication interfaces such as the base stations 230 and 231 in order for providing access for the different devices to the network, and the base stations 230, 231 are themselves connected to the mobile network 220 via a fixed connection 276 or a wireless connection 277.
  • servers 240, 241 and 242 each connected to the mobile network 220.
  • Some of the above devices, for example the computers 240, 241 , 242 may be such that they are arranged to make up a connection to the Internet with the communication elements residing in the fixed network 210.
  • the various devices may be connected to the networks 210 and 220 via communication connections such as a fixed connection 270, 271 , 272 and 280 to the internet, a wireless connection 273 to the internet 210, a fixed connection 275 to the mobile network 220, and a wireless connection 278, 279 and 282 to the mobile network 220.
  • the connections 271 -282 are implemented by means of communication interfaces at the respective ends of the communication connection.
  • Fig. 1 b shows devices for the image segmentation video remixing according to an example embodiment.
  • the server 240 contains memory 245, one or more processors 246, 247, and computer program code 248 residing in the memory 245.
  • the different servers 241 , 242, 290 may contain at least these elements for employing functionality relevant to each server.
  • the end-user device 251 contains memory 252, at least one processor 253 and 256, and computer program code 254 residing in the memory 252 for implementing, for example, the image segmentation process.
  • the end-user device may also have one or more cameras 255 and 259 for capturing image data, for example stereo video.
  • the end-user device may also contain one, two or more microphones 257 and 258 for capturing sound.
  • the end user devices may also comprise a screen for viewing single- view, stereoscopic (2-view), or multiview (more-than-2-view) images.
  • the end-user devices may also be connected to video glasses 290 e.g. by means of a communication block 293 able to receive and/or transmit information.
  • the glasses may contain separate eye elements 291 and 292 for the left and right eye. These eye elements may either show a picture for viewing, or they may comprise a shutter functionality e.g. to block every other picture in an alternating manner to provide the two views of three-dimensional picture to the eyes, or they may comprise an orthogonal polarization filter (compared to each other), which, when connected to similar polarization realized on the screen, provide the separate views to the eyes.
  • Stereoscopic or multiview screens may also be autostereoscopic, i.e. the screen may comprise or may be overlaid by an optics arrangement, which results into a different view being perceived by each eye.
  • Single-view, stereoscopic, and multiview screens may also be operationally connected to viewer tracking such a manner that the displayed views depend on viewer's position, distance, and/or direction of gaze relative to the screen.
  • various processes of image segmentation may be carried out in one or more processing devices; for example, entirely in one user device like 250, 251 or 260, or in one server device 240, 241 , 242 or 290, or across multiple user devices 250, 251 , 260 or across multiple network devices 240, 241 , 242, 290, or across both user devices 250, 251 , 260 and network devices 240, 241 , 242, 290.
  • the elements of the image segmentation process may be implemented as a software component residing on one device or distributed across several devices, as mentioned above, for example so that the devices form a so-called cloud.
  • An embodiment relates to a method for interactive, dynamic and realtime image segmentation usable in data processing devices, especially in touch screen devices, which method enables to effectively select a foreground object from background.
  • a user of the touch screen device is prompted to select the foreground object by providing scribbles on the desired object through the touch screen.
  • the method is based on an algorithm, which uses pixel data located in the neighbourhood of the input scribble and rest of the image, and performs a segmentation to separate the foreground object from background.
  • the algorithm iteratively updates the model parameters used in segmentation during a new stroke information provided by the user and updates the segmented object on the screen.
  • the segmentation method may utilize iterative and dynamic usage of color Gaussian Mixture Model (GMM) and graph cut.
  • GMM color Gaussian Mixture Model
  • Over-segments also referred to as super pixels or turbo pixels in certain processes, are created by grouping similarly colored pixels via merging.
  • the algorithm underlying the method is executed at the over- segmentation level. Therefore, the method is started with segmenting (200) an image into over-segments such that the graph structure and the spatial neighbourhood of the pixel data of the image are preserved.
  • a computationally efficient over-segmentation method is disclosed in "Efficient graph-based image segmentation via speeded-up turbo pixels", by C. Cigla and A.A. Alatan; In Image Processing (ICIP), 2010 17th IEEE International Conference on, pp. 3013 -3016, Sept. 2010.
  • every node represents a over-segment (i.e. super pixel) and there is an edge link between each connected over-segments.
  • the user interaction is started (202) by detecting a user command on the user interface.
  • the user command may be, for example, a tap by a finger or a stylus on a touch screen display, or a click by a mouse or other pointing means on the display area, if a conventional display without a touch input feature is used.
  • the algorithm is arranged to interpret the user command, such as a touch of the user, on a point of the screen as a touch of a brush being centered on said point of the screen and having a predefined radius.
  • the interaction point i.e., the circle-shaped region touched by the brush
  • the brush map is displayed on the original image, and the user interaction is preferably indicated on the display by changing the color of the interaction point.
  • the brush map is updated (204) on the screen at each interaction point.
  • a color GMM is updated (206) at each interaction point.
  • the segmentation process should be started as soon as possible. While the user continues the interaction by providing further user commands (208), the number of interaction points is continuously updated. If the number of interaction points exceeds a predefined threshold (210), the segmentation process (212) is also started.
  • the value of the predefined threshold may vary depending on the properties of the device or the image to be segmented, for example. Generally, a value between 5 and 20 pixels/super pixels, such as 10 pixels/super pixels, can be used as the value of the predefined threshold, but naturally values beyond this range are applicable, as well.
  • the segmentation may be carried out according to the so-called “Grabcut” method, which is disclosed more in detail in “Grabcut: Interactive foreground extraction using iterated graph cuts” by C. Rother, V. Kolmogorov, and A. Blake; ACM Transactions on Graphics, 23:309-314, 2004.
  • a so-called Gibbs energy function E is defined so that its minimum should correspond to a good segmentation, in the sense that it is guided both by the observed foreground and background color GMMs, similarity of pixel-wise GMM components and that the opacity is coherent, reflecting a tendency to solidity of objects.
  • the Grabcut process is carried out iteratively, each iteration step minimising the total energy E with respect to the variables describing pixel-wise GMM components and the opacity of either a foreground or a background object, in turn.
  • E decreases monotonically, thus guaranteeing convergence at least to a local minimum of E.
  • E ceases to decrease significantly, the iteration can be terminated automatically.
  • the Grabcut method can be started with very minimum amount of user interaction. Further, although the entire image is segmented at each iteration step, only the selected foreground object is masked with the brush map and only the foreground region around the brushed area is shown to user. Accordingly, the segmentation may discard the background part of the brushed foreground object and focus only on finding intersection of the brushed region and the object of interest, thus further simplifying the segmentation process. From the usability point of view, it may be beneficial that the user may stop brushing the scribbles on the foreground object at any point, but the algorithm automatically continues the segmentation on the basis of the user interaction provided so far.
  • a spatio-color model can be used instead of using a color GMM based background and foreground model.
  • a spatio-color model can be used instead of the GMM.
  • a color histogram or an intensity histogram can be used to model the foreground and the background.
  • any clustering or classification algorithm can be used to define a model which distinguishes foreground and background.
  • the likelihood of foreground and background GMM for finding edge costs of the graph
  • some other method can also be used, like determining the distance to a mean color.
  • a method based on alpha/beta swapping may also be used to segment the image into background and foreground. Any other energy minimization, such as simulated annealing or belief propagation, may also be used to minimize the defined cost function.
  • the user interaction with the image on the display may be arranged to simulate a coloring gesture to make interaction entertaining and intuitive. Initially, a grayscale version of the image is provided on the display and the user start to color an object of interest within the image.
  • the result of the segmentation is shown to user by coloring the foreground object; i.e. the pixels belonging to foreground object is shown in color, whereas background pixels are shown in grayscale, as shown in Figure 3a.
  • the user uses this feedback from the device, the user continues drawing scribbles and colors the remaining areas of the foreground object, as shown in Figures 3b and 3c.
  • a model is learned throughout the interaction. When the interaction continues and if a new interaction point is not consistent with this model, thus suggesting that the point is outside the boundaries of the object of interest, that point is recorded as an exit point. Then, if a later interaction point provided by the user enters back to a region, which is consistent with the previous model, that point is recorded as an entrance point. Then a correct path between the exit and the entrance points is determined by local search.
  • FIG 4 This is illustrated in Figure 4, where the user has started drawing scribbles on the foreground object according to path 400.
  • the scribbles are compared to the underlying object of interest, and a model of the presumed intention of the user is created on the basis of this comparison.
  • the user maybe accidentally, draws scribbles outside the boundaries of the object of interest. This is not consistent with the created model, and the point 402 is defined as an exit point.
  • the user draws scribbles towards the foreground object and enters the boundaries of the object of interest at point 404.
  • point 404 is defined as an entrance point.
  • the algorithm defines a correct path between the exit point 402 and the entrance point 404 along the boundary of the object of interest, and inserts the correct path in the brush map. Later along the path of scribbles, a similar exit occurs at point 406 and an entrance at point 408. Again, a correct path is defined between the exit point 406 and the entrance point 408 along the boundary of the object of interest, and the correct path is inserted in the brush map.
  • the brush size around the user interaction point i.e. the input scribble
  • the brush size around the user interaction point is automatically adjusted based on the local image characteristics around the input scribble. Based on the obtained local texture, thick brush size is assigned to homogenous regions and thin brush size is assigned to regions with high textures.
  • the above method is applicable to interactive video segmentation, as well.
  • the interaction of the user in the first frame may be utilised in the video segmentation such that the interaction is propagated through the subsequent frames.
  • the user interaction is started by drawing scribbles on a foreground object in the first frame.
  • the user interaction is continued in the first frame, for example as described in Figure 3, whereafter the algorithm as described above solves the minimization problem and shows the segmented boundary on the screen.
  • the segmentation algorithm uses the over-segmented regions (super pixels), the previous interaction result is utilized in terms of those super pixels.
  • the foreground and background samples of the first frame are assumed to be known.
  • over- segmented regions are used to estimate the super pixels, which are closer to the previously selected super pixels in terms of a similarity metric.
  • the propagation of the interaction result through the video sequence to segment the subsequent frames is accomplished by using a search algorithm on the super pixels, not using any motion estimation algorithm.
  • Segmentation of a foreground region in a video sequence is based on a color model obtained from the interaction result of the previous frames. Color models are extracted from the interaction with an initialization taken from the previous frames. Such an approach helps to make the color model accurate and fast to acquire.
  • the regions are given to be modeled with the Gaussian Mixture. Except the first frame, the previous GMM information is provided as initials of the Expectation Maximization algorithm, which is carried out to find GMM for the given interaction. By performing such an initialization, the convergence and accuracy increases and becomes faster.
  • the calculated Gaussian Mixture Model is used in constructing the graph as in the still image part. Finally the segmentation is performed using a graph cut algorithm.
  • Figures 6a and 6b illustrate the propagation of the user interaction from a first frame to a second frame.
  • the user draws a scribble on the human figure considered a foreground object.
  • the super pixels of the scribble of Figure 6a are used as a basis for estimating the super pixels in the next frame ( Figure 6b), which are sufficiently similar to the super pixels selected in the first frame.
  • the resulting super pixels are shown as a lighter color scribble in Figure 6b.
  • the super pixels of this scribble are used as a basis for estimating the super pixels in the next frame, and so on.
  • no computationally heavy motion estimation is used for segmenting the foreground objects, but only a light-weight similarity search is carried out for the super pixels.
  • the various embodiments may provide advantages over state of the art. With rather minimum amount of user interaction, an accurate and pleasant looking segmentation and 3D perception may be achieved.
  • the various embodiments provide real-time image segmentation, which is remarkably robust to interaction errors. From the usability point of view, the overall process is intuitive and entertaining for the user. Furthermore, the process is adaptive to complicated textures of foreground objects.
  • a terminal device may comprise circuitry and electronics for handling, receiving and transmitting data, computer program code in a memory, and a processor that, when running the computer program code, causes the terminal device to carry out the features of an embodiment.
  • a network device may comprise circuitry and electronics for handling, receiving and transmitting data, computer program code in a memory, and a processor that, when running the computer program code, causes the network device to carry out the features of an embodiment.
  • the various devices may be or may comprise encoders, decoders and transcoders, packetizers and depacketizers, and transmitters and receivers.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Probability & Statistics with Applications (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Processing Or Creating Images (AREA)

Abstract

Un procédé et un appareil conçu pour fournir une image sur un affichage ; détecter une commande d'utilisateur sur l'affichage, la commande de l'utilisateur sélectionnant une région de l'image sous la forme d'un objet de premier plan ; mettre à jour une image de masque au niveau d'un emplacement de la commande de l'utilisateur ; actualiser un algorithme de classification distinguant l'objet d'avant-plan et une région d'arrière-plan de l'image ; répéter les étapes en commençant à partir de ladite détection jusqu'à ce qu'un nombre prédéfini de commandes d'utilisateur a été détecté ; et démarrer un processus de segmentation de l' image.
PCT/FI2012/050313 2012-03-29 2012-03-29 Segmentation d'image WO2013144418A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/FI2012/050313 WO2013144418A1 (fr) 2012-03-29 2012-03-29 Segmentation d'image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/FI2012/050313 WO2013144418A1 (fr) 2012-03-29 2012-03-29 Segmentation d'image

Publications (1)

Publication Number Publication Date
WO2013144418A1 true WO2013144418A1 (fr) 2013-10-03

Family

ID=49258297

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/FI2012/050313 WO2013144418A1 (fr) 2012-03-29 2012-03-29 Segmentation d'image

Country Status (1)

Country Link
WO (1) WO2013144418A1 (fr)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2846309A1 (fr) * 2013-08-27 2015-03-11 Samsung Electronics Co., Ltd Procédé et appareil de segmentation d'objet dans une image
US9542751B2 (en) 2015-05-08 2017-01-10 Qualcomm Incorporated Systems and methods for reducing a plurality of bounding regions
US9865062B2 (en) 2016-02-12 2018-01-09 Qualcomm Incorporated Systems and methods for determining a region in an image
US9886767B2 (en) 2013-05-03 2018-02-06 Nokia Technologies Oy Method, apparatus and computer program product for segmentation of objects in images
US10620826B2 (en) 2014-08-28 2020-04-14 Qualcomm Incorporated Object selection based on region of interest fusion
CN111582060A (zh) * 2020-04-20 2020-08-25 浙江大华技术股份有限公司 自动划线周界报警方法、计算机设备及存储装置
CN112345450A (zh) * 2020-10-29 2021-02-09 钢研纳克检测技术股份有限公司 大尺寸不规则样品表面扫描区域识别及扫描路径确定方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080136820A1 (en) * 2006-10-20 2008-06-12 Microsoft Corporation Progressive cut: interactive object segmentation
US20090304280A1 (en) * 2006-07-25 2009-12-10 Humaneyes Technologies Ltd. Interactive Segmentation of Images With Single Scribbles
US20110216976A1 (en) * 2010-03-05 2011-09-08 Microsoft Corporation Updating Image Segmentation Following User Input
US20110243443A1 (en) * 2008-12-09 2011-10-06 Koninklijke Philips Electronics N.V. Image segmentation

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090304280A1 (en) * 2006-07-25 2009-12-10 Humaneyes Technologies Ltd. Interactive Segmentation of Images With Single Scribbles
US20080136820A1 (en) * 2006-10-20 2008-06-12 Microsoft Corporation Progressive cut: interactive object segmentation
US20110243443A1 (en) * 2008-12-09 2011-10-06 Koninklijke Philips Electronics N.V. Image segmentation
US20110216976A1 (en) * 2010-03-05 2011-09-08 Microsoft Corporation Updating Image Segmentation Following User Input

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
ROTHER, C. ET AL.: "GrabCut - interactive foreground extraction using iterated graph cuts.", ACM TRANSACTIONS ON GRAPHICS (TOG) - PROCEEDINGS OF ACM SIGGRAPH 2004, vol. 23, no. 3, August 2004 (2004-08-01), pages 309 - 314, XP002340109, Retrieved from the Internet <URL:http://dl.acm.org/citation.cfm?doid=1015706.1015720> [retrieved on 20130129], DOI: doi:10.1145/1015706.1015720 *
TASLI, H. ET AL.: "Interactive 2D 3D image conversion method for mobile devices.", 3DTV CONFERENCE: THE TRUE VISION - CAPTURE, TRANSMISSION AND DISPLAY OF 3D VIDEO (3DTV-CON), 16 May 2011 (2011-05-16), Retrieved from the Internet <URL:http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5877212> [retrieved on 20130129] *
WANG, J. ET AL.: "Soft scissors: an interactive tool for realtime high quality matting.", ACM TRANSACTIONS ON GRAPHICS (TOG) - PROCEEDINGS OF ACM SIGGRAPH 2007, vol. 26, no. 3, July 2007 (2007-07-01), Retrieved from the Internet <URL:http://dl.acm.org/citation.cfm?id=1276389> [retrieved on 20130129] *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9886767B2 (en) 2013-05-03 2018-02-06 Nokia Technologies Oy Method, apparatus and computer program product for segmentation of objects in images
EP2846309A1 (fr) * 2013-08-27 2015-03-11 Samsung Electronics Co., Ltd Procédé et appareil de segmentation d'objet dans une image
US9478040B2 (en) 2013-08-27 2016-10-25 Samsung Electronics Co., Ltd Method and apparatus for segmenting object in image
US10235761B2 (en) 2013-08-27 2019-03-19 Samsung Electronics Co., Ld. Method and apparatus for segmenting object in image
US10620826B2 (en) 2014-08-28 2020-04-14 Qualcomm Incorporated Object selection based on region of interest fusion
US9542751B2 (en) 2015-05-08 2017-01-10 Qualcomm Incorporated Systems and methods for reducing a plurality of bounding regions
US9865062B2 (en) 2016-02-12 2018-01-09 Qualcomm Incorporated Systems and methods for determining a region in an image
CN111582060A (zh) * 2020-04-20 2020-08-25 浙江大华技术股份有限公司 自动划线周界报警方法、计算机设备及存储装置
CN111582060B (zh) * 2020-04-20 2023-04-18 浙江大华技术股份有限公司 自动划线周界报警方法、计算机设备及存储装置
CN112345450A (zh) * 2020-10-29 2021-02-09 钢研纳克检测技术股份有限公司 大尺寸不规则样品表面扫描区域识别及扫描路径确定方法
CN112345450B (zh) * 2020-10-29 2023-10-13 钢研纳克检测技术股份有限公司 大尺寸不规则样品表面扫描区域识别及扫描路径确定方法

Similar Documents

Publication Publication Date Title
Valentin et al. Depth from motion for smartphone AR
JP6655737B2 (ja) マルチビューシーンのセグメンテーションおよび伝播
US9237330B2 (en) Forming a stereoscopic video
US9041819B2 (en) Method for stabilizing a digital video
US20130127988A1 (en) Modifying the viewpoint of a digital image
US9495755B2 (en) Apparatus, a method and a computer program for image processing
WO2016110239A1 (fr) Procédé et dispositif de traitement d&#39;image
US8610712B2 (en) Object selection in stereo image pairs
WO2013144418A1 (fr) Segmentation d&#39;image
US8611642B2 (en) Forming a steroscopic image using range map
US10140513B2 (en) Reference image slicing
US20130129192A1 (en) Range map determination for a video frame
KR101401184B1 (ko) 동영상의 객체 경계 추정 방법
US9165401B1 (en) Multi-perspective stereoscopy from light fields
CN109564691B (zh) 用于基于局部确定的3d信息假说执行3d估计的装置和方法
Luo et al. Foreground removal approach for hole filling in 3D video and FVV synthesis
WO2015188666A1 (fr) Procédé et dispositif de filtrage vidéo en trois dimensions
Yang et al. Dynamic 3D scene depth reconstruction via optical flow field rectification
US9959672B2 (en) Color-based dynamic sub-division to generate 3D mesh
Hsu et al. Spatio-temporally consistent view synthesis from video-plus-depth data with global optimization
US9786055B1 (en) Method and apparatus for real-time matting using local color estimation and propagation
He et al. Iterative transductive learning for automatic image segmentation and matting with RGB-D data
Iizuka et al. Efficient depth propagation for constructing a layered depth image from a single image
US9171357B2 (en) Method, apparatus and computer-readable recording medium for refocusing photographed image
Lee et al. Estimating scene-oriented pseudo depth with pictorial depth cues

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 12872539

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 12872539

Country of ref document: EP

Kind code of ref document: A1