EP2353163A1 - Procede de decoupage de contenu multimedia, dispositif et programme d'ordinateur correspondant - Google Patents
Procede de decoupage de contenu multimedia, dispositif et programme d'ordinateur correspondantInfo
- Publication number
- EP2353163A1 EP2353163A1 EP09760211A EP09760211A EP2353163A1 EP 2353163 A1 EP2353163 A1 EP 2353163A1 EP 09760211 A EP09760211 A EP 09760211A EP 09760211 A EP09760211 A EP 09760211A EP 2353163 A1 EP2353163 A1 EP 2353163A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- scenes
- cutting
- tree
- chapters
- chapter
- Prior art date
- Legal status (The legal status 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 status listed.)
- Ceased
Links
- 238000005520 cutting process Methods 0.000 title claims abstract description 74
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000004590 computer program Methods 0.000 title claims description 5
- 230000002123 temporal effect Effects 0.000 claims description 24
- 230000011218 segmentation Effects 0.000 claims description 16
- 238000004458 analytical method Methods 0.000 claims description 15
- 238000000638 solvent extraction Methods 0.000 claims description 14
- 238000010276 construction Methods 0.000 claims description 12
- 238000002844 melting Methods 0.000 claims description 3
- 230000008018 melting Effects 0.000 claims description 3
- 238000005192 partition Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims 1
- 230000000717 retained effect Effects 0.000 description 7
- 238000000605 extraction Methods 0.000 description 6
- 238000012512 characterization method Methods 0.000 description 4
- 239000008186 active pharmaceutical agent Substances 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 101150012579 ADSL gene Proteins 0.000 description 1
- 102100020775 Adenylosuccinate lyase Human genes 0.000 description 1
- 108700040193 Adenylosuccinate lyases Proteins 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
- G11B27/00—Editing; Indexing; Addressing; Timing or synchronising; Monitoring; Measuring tape travel
- G11B27/10—Indexing; Addressing; Timing or synchronising; Measuring tape travel
- G11B27/19—Indexing; Addressing; Timing or synchronising; Measuring tape travel by using information detectable on the record carrier
- G11B27/28—Indexing; Addressing; Timing or synchronising; Measuring tape travel by using information detectable on the record carrier by using information signals recorded by the same method as the main recording
Definitions
- a method of dividing multimedia content, device and corresponding computer program A method of dividing multimedia content, device and corresponding computer program.
- the present invention relates to the field of chapter definition for multimedia content.
- the present invention relates more particularly to a method of chapitrage of multimedia contents which comprise at least one video stream.
- the adaptation of these contents takes several aspects: - an adaptation of the image resolution, aspect ratio width / height; transcoding to mobile formats (3GP); a cutting of the original video of multimedia contents in short chapters.
- the invention relates more particularly to this last point.
- the division of a video into chapters also called “chapters” makes it possible to adapt to the "mobile" consumption of multimedia content, consisting of short content particularly adapted for viewing in a mobile situation. It has indeed been found that the viewing of a multimedia content on a mobile terminal did not exceed, in general, the ten minutes. This is particularly because the screens of mobile terminals are relatively small. Also, for ease of viewing, the long contents are divided into segments (chapters) of five to ten minutes which facilitates the resumption of viewing of video content after an interruption. 2 SOLUTIONS OF THE PRIOR ART
- Time constraints dictate that the division must produce files (chapters) of homogeneous duration. For example, a time constraint may require that files be between five and ten minutes long. A second temporal constraint may require that each duration, taken independently, can not differ by more than one minute from the average duration of the files produced.
- Semantic constraints also called editorial constraints
- the division should preferably be done on a change of plans, at the ends of a scene, the scene being the elementary semantic unit resulting from the concatenation of plans forming a unit, taking place in the same place at the same time (each plan is characterized by a continuity of filmed content, so without cut or change of view).
- a dialogue, a telephone conversation, an action observed from several points of view are scenes that we do not wish to cut at the time of division into chapters It is therefore understood that the manual creation of the cutting of files, which takes into account both temporal constraints and editorial constraints is long, tedious and expensive
- the generally observed approach is thus of the "bottom-up" type, starting from the elementary entities that are the plans to define groups of similar plans then scenes.
- the result obtained by these techniques is very dependent on characteristics extracted from the audio / video stream aiming to group the shots, of mixing criteria when several characteristics are used to group the shots (color, texture of the picture, power and sound frequencies). , the granula ⁇ té chosen to form clusters (or groupings) of plans, criteria retained to merge clusters into scenes.
- These techniques require the use of several comparison thresholds having a direct impact on the result and creating scenes more or less respectful of the desired semantic division.
- the invention overcomes these disadvantages of the prior art.
- the invention relates in fact to a method of cutting, in at least two chapters, a multimedia content comprising at least one video stream,
- such a method comprises: a segmentation phase of said at least one video stream into a plurality of plans defining scenes; a phase of analysis of similarities between scenes, associating at least one proximity information with at least two scenes considered; a phase of constructing chapters of said video stream according to said proximity information of said scenes and at least one predetermined time division constraint, which defines at least one information characterizing the duration of a chapter; a construction phase of said chapters, according to said cutting hypotheses.
- said resemblance analysis phase comprises a step of constructing a resemblance tree comprising a hierarchy of scenes constructed by merging similar scenes; said phase of constructing chapter partitioning hypotheses comprises a step of creating at least one global hypothesis of cutting said multimedia content according to said resemblance tree and according to said at least one predetermined time-division constraint, said at least one at least one global partitioning hypothesis comprising a plurality of local chapter splitting assumptions; and said phase of constructing said chapters includes a step of selecting at least one global partitioning hypothesis from among said constructed partitioning assumptions.
- the invention makes it possible to respect both temporal constraints and editorial constraints by constructing a hierarchical data structure that iteratively associates scenes with each other to form higher level scenes.
- said at least one predetermined time-division constraint comprises at least a maximum duration of a chapter of said multimedia content and / or at least a minimum duration of a chapter of said multimedia content.
- said cutting method further comprises a verification phase of the validity of said at least one overall hypothesis of chapters of said video stream by relating to a segmentation of at least one audio stream of said multimedia content.
- the invention makes it possible to validate the global switching hypotheses as a function of the audio stream that accompanies the video stream, to ensure that the chapters that have been defined are coherent with the audio stream of the content.
- the invention ensures that chapters defined using the resemblance structure do not cut a dialogue or a speech flow in an untimely manner.
- said step of constructing said resemblance tree comprises: a step of determining at least one descriptor for at least one plane of said plurality of planes; a step of calculating a distance between the descriptors of at least two planes; a step of melting said planes having a minimum distance, forming a scene; a step of inserting said scene into said resemblance tree.
- the invention makes it possible to group the shots of the video stream into a first level of scenes.
- These first stages of construction of the resemblance tree allow, from the plans, which constitute elementary scenes to define a first level of the hierarchy of scenes in the resemblance tree.
- said step of determining the descriptors of said planes of said plurality of planes comprises, for a current plan: a subsampling step of said current plan in a plurality of representative images of said current plan; a step of calculating at least one descriptor for at least some of said images of said plurality of images; a step of merging said at least one calculated descriptor, delivering a descriptor of said current plan.
- the invention realizes a characterization of a plane based on several images representative of the plane and not on a single image of the plane.
- said step of merging said planes takes into account a maximum duration between two plans to be merged.
- the invention makes it possible to ensure that each scene formed from planes having a minimum distance does not exceed a predetermined duration.
- This maximum duration can prevent the merger of two plans that are separated by too long a distance.
- this parameter avoids, in an extreme case, that two planes that are located at both ends of the video stream are merged. In a more conventional manner, this parameter avoids the merger of plans that would for example be 10 minutes apart, which could correspond to the maximum length of a chapter.
- the scenes directly from the planes are constructed according to a maximum duration between two plans to merge which can for example be of the order of the minute, so that each of the elementary scenes defined in the resemblance tree respects constraints of proximities of the planes and a constraint of duration.
- said descriptor is representative of a colo ⁇ mét ⁇ que distribution of a plurality of images composing a plane.
- the invention implements a simple colo ⁇ metric distribution, which does not take into account the spatiality characteristics and is therefore suitable for calculating the distance between planes for the invention.
- the color distribution is particularly suitable because it relates different frames of the same physical scene (eg the different frames of a dialogue scene in the same scenery).
- said step of constructing said resemblance tree comprises at least one iteration of the following steps calculating a distance between at least some of said scenes of said resemblance tree, - grouping scenes of said plurality of scenes within said resemblance tree, according to said distances of scenes computed so that scenes of a lower hierarchical level are grouped together in a scene of a higher hierarchical level in said tree
- the invention constructs a tree in which scenes of a lower level are used to define the scenes of the higher level by comparing the distances separating these scenes and grouping the scenes of the lower level according to these distances.
- a distance calculation function which can for example take into account the descriptors of the planes that make up the scenes whose distance is to be calculated.
- said step of constructing said resemblance tree ends when said scene grouping step delivers a single scene for the entire video stream
- the process of creating the resemblance tree stops when a single scene is obtained by grouping scenes of the lower level.
- the resemblance tree comprises a plurality of scenes. in which the planes constitute the basic unit for obtaining the scenes.
- the invention makes it possible to construct a resemblance tree which makes it possible to associate the plans with each other in order to construct scenes and to associate, iteratively. or recursively these scenes together to form a set of scenes that are hierarchically grouped within the likeness tree
- said phase of building chapter splitting hypotheses comprises - a step of searching, within the tree of similarity, of a plurality local assumptions of cutting said flow, said local breakdown assumptions unitarily respecting said at least one temporal cutting constraint; a step of combining said local partitioning assumptions forming said at least one global partitioning hypothesis; a step of inserting said at least one global partition hypothesis into a list of global hypotheses grouping said global hypotheses.
- the invention makes it possible to construct global partitioning hypotheses, that is to say a set of candidate chapters, which may be suitable for the division of the multimedia content, according to the local hypotheses obtained thanks firstly to the grouping of scenes and secondly to the temporal constraint of cutting.
- Global assumptions can be based on very variable local assumptions of cutting. Thus, some global hypotheses in the list can propose cuts of the flow in three parts whereas others can propose cuts in five parts.
- the selection of the global hypothesis chosen for the cut can either be done after an audio segmentation that allows the hypotheses to be compared to moments of speech and non-speech.
- the overall hypothesis retained can also be selected according to other criteria such as a minimum number of chapters or the average distance separating the local hypotheses composing the global hypothesis selected or a combination of such criteria.
- the invention also relates to a device for chapter-splitting a multimedia content comprising at least one video stream.
- such a device comprises: means for segmenting said at least one video stream into a plurality of plans defining scenes; means for analyzing similarities between scenes, associating the less a proximity information to at least two scenes considered; means for constructing chapters of said video stream as a function of said proximity information of said scenes and at least one predetermined time-division constraint, which defines at least one information characterizing the duration of a chapter; means for constructing said chapters, according to said cutting hypotheses.
- Such a device may also contain complementary means for implementing the cutting method as described above.
- the invention also relates to a computer program comprising instructions for implementing the cutting method described above when this program is executed by a processor, or a computer program product downloadable from a communication network and / or stored on a computer-readable and / or executable medium by a processor, comprising program code instructions for implementing the chopping method described above, when executed by a processor. 4 LIST OF FIGURES
- FIG. 1 presents a division of a video stream into a plane
- FIG. 2A illustrates a first embodiment of the cutting according to the invention
- FIG. 2B illustrates a second embodiment of cutting according to the invention
- FIG. 3 describes a method for calculating a plan descriptor according to the invention
- - Figure 4 presents the method of calculating the resemblance tree according to the invention
- FIG. 5 illustrates a similarity tree obtained by grouping similar scenes according to the invention
- FIG. 6 presents cutting hypotheses using the similarity tree of FIG. 5.
- the general principle of the invention is based on taking into account the temporal constraint for the creation of the chapters of the multimedia content.
- This temporal constraint is taken into account by realizing, on the one hand, hypotheses of cutting the multimedia content according to a particular method, taking into account a maximum duration of separation in scenes.
- the temporal constraint is also taken into account by performing a verification on the other hand to determine, among several given partitioning assumptions, which ones best respect the temporal constraint.
- the invention therefore defines a method of cutting, in at least two chapters, a multimedia content (CM) comprising at least one video stream
- CM multimedia content
- the method of the invention comprises a segmentation phase (FIG. 2A, 201) of the video stream in a plurality of planes.
- This segmentation phase (FIG. 2A, 201) makes it possible, as described in FIG. 1, to partition the video stream (FV) in the plane P / to Pn, each of these planes comprising a set of images.
- the method of the invention also comprises: a phase of analysis of resemblances between the planes which associates proximity information between the plans considered (FIG.
- This temporal cutting constraint defines at least one piece of information that characterizes the duration of a chapter.
- the method makes it possible to obtain one or more chapter splitting results (RES).
- the similarity analysis phase (202) comprises a step of constructing a resemblance tree, which is a hierarchical structure that makes it possible, firstly, to group the plans into one another. function of proximity information and secondly to iteratively build the assumptions of grouping plans.
- This construction of the plan clustering assumptions includes a step of creating at least one global hypothesis of cutting the multimedia content according to the similarity tree.
- a global slicing hypothesis comprises, according to the invention, a plurality of local hypotheses for cutting multimedia content into chapters.
- the invention hierarchically represents the plans and scenes of the video of the multimedia content.
- the possibilities of cutting the video are as numerous as the possibilities of partitioning the tree into several sub-trees.
- each leaf of the tree represents a plane and each node grouping two sheets or more nodes in the higher levels represents a "possible scene”.
- the invention makes it possible to approach this behavior by implementing a hierarchical classification algorithm to reveal the similarities between temporally close planes that most often belong to the same scene. These similarities are represented in the lowest level of the tree. These similarities are detected using the proximity information between plans. To determine the proximity between plans and to construct the classification of these plans between them, the inventors chose to use only one descriptor (the color distribution) and only one modality (the video). Such an approach makes it possible to avoid introducing parameters or thresholds to be set a priori (no mixture of characteristics).
- this color distribution is generally not dependent on the spatial evolution of the plane: for example if a camera moves in space while always aiming for the same character, the colorimetric distribution of the plan does not evolve much, to the difference of a spatial distribution for example.
- the color distribution is therefore particularly well suited to the first phase of grouping shots in order to constitute possible scenes.
- such a distribution is simple to implement.
- Other types of proximity determination functions are of course conceivable as long as they make it possible to establish a grouping of shots into scenes and then a grouping of scenes into other higher-ranking scenes.
- a mixed analysis based on the video on the one hand and on the audio on the other hand is however implemented.
- the audio analysis makes it possible to detect the scenes of dialogues not to be cut in the middle of a dialogue.
- the inventors have therefore placed an audio analysis (206) at the end of the validation process of the global cutting hypotheses.
- This audio analysis has the function of filtering the hypotheses extracted from the video analysis which cut the multimedia content into chapters which certainly respect the temporal constraints, but cut dialogue scenes in an impromptu way.
- the solution adopted to obtain cutting assumptions is therefore based on two prior segmentation of the contents • a video segmentation (201) cutting the content of video clips (detection of sharp transitions or progressive); the audio segmentation (205) generates, by spectral analysis, segments of three types: silence, speech or no word.
- the tree of resemblance is a hierarchical data structure in which the plans resulting from the segmentation constitute the leaves of the tree. These sheets will be successively used to build a tree of scenes, each scene of higher rank in the tree being made up of at least two scenes of lower rank, and this according to proximity information.
- the invention proceeds before cutting the video into plans. This division is performed according to the methods of the state of the art by detection of sudden or progressive transitions. A set of plans is thus obtained, as shown in Figure 3 (plans Pi, Pj and Pk). These plans constitute the basic scenes of the tree of resemblance. The plans are therefore elementary scenes.
- each image descriptor is merged into a global descriptor in the plane: H 1 and H J in the example of Figure 3.
- Hi is the descriptor of the plane Pj
- Hj is the descriptor Plan Pj.
- This histogram is calculated on each sub-sampled image of the plane (the images II to In, for example in FIG. 3), the fusion being carried out by summation of all the histograms resulting from the selected images and then normalization so as to make comparable the long and short shots (for which the number of images would be different).
- the value D must be chosen sufficiently large to be greater than all Chi-square distances calculated for ⁇ T J ⁇ T _ In practice, a value of 10 6 , or the maximum value representable by the variable may be used. In this embodiment, the calculated distances are inserted in a triangular matrix. These distances thus serve as starting points for the construction of the similarity tree according to the invention.
- DS is the distance between the scene S 1 and the scene S J ;
- T represents the set of planes belonging to the scene S ';
- r represents the set of planes belonging to the scene S J ;
- LJ represents the distance between the plane k of the scene S and the plane / of the scene S J.
- the distance between two scenes is considered as the minimum distance between the respective planes of these two scenes.
- a first step will thus consist in calculate a first series of possible scenes, of rank immediately superior to the planes (which constitute the leaves of the similarity tree according to the invention).
- the following steps will consist in successively calculating, from the lower-ranked scenes, scenes of the next higher rank
- N histograms denoted H are calculated for each of the N planes of the video
- a list L of scenes (401), initially containing N elements is created: each scene is composed of a single plane, resulting from the initial segmentation of the video. The scenes that make up this list constitute the leaves of the tree.
- the distance DS * ' the minimum distance between all the pairs (i, j) among the N scenes of L is calculated.
- a new scene S N connecting S k to S 1 (402) is added to the list L of scenes for the minimum distance found.
- the scenes S * and S 1 are removed from L Then the number of scenes of 1 is incremented to form a new scene identifier;
- step 4 is repeated until list L contains only one scene
- the next phase according to the invention consists in extracting from this tree hypotheses of cutting the video stream based precisely on the distances that separate the scenes.
- the search and extraction of hypotheses are carried out by traversing the tree from top to bottom.
- the scenes with the highest distance are searched in priority to build the cutting hypotheses.
- An important parameter of the extraction is the taking into account of the temporal constraint, which makes it possible to determine a maximum duration and a minimum duration of a chapter.
- the search and the extraction of hypotheses are carried out according to the following parameters: the duration of the video is D ID ; the maximum duration of chapters is D MAX ; the minimum duration of the chapters is fixed at D M ⁇ X / 2 to avoid the non-feasibility of the constraints: thus one ensures to always find a suitable solution to the constraints; a maximum variation around the average, D V ⁇ R is defined. It makes it possible to ensure a homogeneity constraint of duration between the chapters.
- the extraction phase of the cutting hypotheses presented in relation to FIGS. 5 and 6 is as follows: 1. the chapter duration constraint [D MAX / 2, D MAX ] determines the validity interval for the choice of cutting assumptions (in dotted line in Figure 5). 2. The descending run from the node allows to define the best cutting points: Figure 6 shows how the best score hypothesis is chosen: for each node, the min and max times are extracted from the boundaries of all the nodes. plans lying under the considered node
- the path above may be iterated several times, so as to extract several local assumptions cutting, so that the 1 st hypothesis has the strongest score, the second having the next lower score, and so on .
- each local hypothesis on the 1st chapter involves a new range of validity to the end of the second chapter, and thus new local assumptions for the end of the second chapter. So for 5 local assumptions 1 chapter 25 local hypotheses will be evaluated for the second chapter (5 new local hypotheses cut for the second chapter for each of the 5 local assumptions cutting 1 chapter). Local hypotheses are thus combined to form global hypotheses of divisions. These combinations are not random, however, since a given local hypothesis can not be combined with any other. It is necessarily combined with a local hypothesis which allowed to assume a division of a first chapter.
- each global hypothesis including a variable number of chapters 6 chapters of 5 mn or 5 chapters of 6 mn are potentially possible
- an average score (average of the scores of each cut) 8.
- the list of global hypotheses is sorted according to the mean scores of the hypotheses • the highest average score corresponding to the best sequence of cuts
- each global hypothesis score being the average of the average distances of the scenes which compose each local hypothesis of cutting
- the set of global hypothesis lists is compared to the audio segmentation.
- This audio segmentation differentiates three types of temporal segments silences, speech segments and those of non-speech
- a list of local hypotheses (that is to say a global hypothesis) is validated if the local hypotheses of cutting constituting it are located in segments of non-speech type, of a duration greater than TNP (parameter of the 'application). Providing that the non-speech time within which a cut is allowed must be greater than the parameter TNP (for example 10 seconds) makes it possible to be sure not to cut in the middle of a dialogue.
Landscapes
- Television Signal Processing For Recording (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR0857722 | 2008-11-13 | ||
| PCT/FR2009/052050 WO2010055242A1 (fr) | 2008-11-13 | 2009-10-26 | Procede de decoupage de contenu multimedia, dispositif et programme d'ordinateur correspondant |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP2353163A1 true EP2353163A1 (fr) | 2011-08-10 |
Family
ID=40474773
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP09760211A Ceased EP2353163A1 (fr) | 2008-11-13 | 2009-10-26 | Procede de decoupage de contenu multimedia, dispositif et programme d'ordinateur correspondant |
Country Status (2)
| Country | Link |
|---|---|
| EP (1) | EP2353163A1 (fr) |
| WO (1) | WO2010055242A1 (fr) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2408190A1 (fr) * | 2010-07-12 | 2012-01-18 | Mitsubishi Electric R&D Centre Europe B.V. | Détection de limites vidéo sémantiques |
| EP2642487A1 (fr) * | 2012-03-23 | 2013-09-25 | Thomson Licensing | Segmentation vidéo à multigranularité personnalisée |
| BR112015021139A2 (pt) * | 2013-03-08 | 2017-07-18 | Thomson Licensing | método e aparelho para segmentação automática de vídeo |
| FR3099674B1 (fr) * | 2019-08-03 | 2022-04-08 | Synchronized | Procede et appareil d’enrichissement de contenu multimedia par des meta-informations |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000045603A1 (fr) * | 1999-01-29 | 2000-08-03 | Sony Corporation | Procede de traitement des signaux et dispositif de traitement de signaux video/vocaux |
| US7224892B2 (en) * | 2001-06-26 | 2007-05-29 | Canon Kabushiki Kaisha | Moving image recording apparatus and method, moving image reproducing apparatus, moving image recording and reproducing method, and programs and storage media |
| US7296231B2 (en) * | 2001-08-09 | 2007-11-13 | Eastman Kodak Company | Video structuring by probabilistic merging of video segments |
-
2009
- 2009-10-26 EP EP09760211A patent/EP2353163A1/fr not_active Ceased
- 2009-10-26 WO PCT/FR2009/052050 patent/WO2010055242A1/fr not_active Ceased
Non-Patent Citations (2)
| Title |
|---|
| None * |
| See also references of WO2010055242A1 * |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2010055242A1 (fr) | 2010-05-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP0406074B1 (fr) | Procédé de segmentation du champ de mouvement d'une image pour le codage d'images vidéo | |
| US9813784B1 (en) | Expanded previously on segments | |
| EP3225027B1 (fr) | Procédé de composition d'une représentation vidéo intermédiaire | |
| EP2353163A1 (fr) | Procede de decoupage de contenu multimedia, dispositif et programme d'ordinateur correspondant | |
| WO2008074877A1 (fr) | Procede de creation d'un nouveau sommaire d'un document audiovisuel comportant deja un sommaire et des reportages et recepteur mettant en œuvre le procede | |
| EP2524324B1 (fr) | Procede de navigation parmi des identificateurs places dans des zones et recepteur mettant en oeuvre le procede | |
| EP3378232B1 (fr) | Procédé de traitement de données codées, procédé de réception de données codées, dispositifs, et programmes d'ordinateurs associés | |
| FR2815151A1 (fr) | Procede de recherche d'images dans un film et dispositif faisant application dudit procede | |
| WO2009081016A1 (fr) | Procede d'analyse d'un contenu multimedia, produit programme d'ordinateur et dispositif d'analyse correspondants | |
| EP3245794A1 (fr) | Procédé de transmission d'un flux de données utilisant un protocole de diffusion en direct | |
| FR3053555A1 (fr) | Procede de codage d'une image numerique, procede de decodage, dispositifs, terminal et programmes d'ordinateurs associes | |
| Furini et al. | On using clustering algorithms to produce video abstracts for the web scenario | |
| EP1182878A1 (fr) | Système de communication, émetteur, récepteur, méthode utilisant un descripteur de stockage de données | |
| EP1343327B1 (fr) | Procédé pour effectuer un traitement sur un contenu multimedia | |
| EP2401700B1 (fr) | Traitement d'un flux de données numériques | |
| WO2018114108A1 (fr) | Procede d'enregistrement d'un programme telediffuse a venir | |
| FR2816793A1 (fr) | Dispositif de traitement d'information multimedia | |
| EP4078959B1 (fr) | Procede d'encodage d'une image numerique en vue de sa compression | |
| EP4564823A1 (fr) | Passerelle pour encodage local de contenus de télévision numérique terrestre en segments de contenus adaptatifs sur http (has) | |
| EP2145499B1 (fr) | Procédé de fourniture de services personnalisés à un terminal par un réseau visité | |
| Gan et al. | A Zero Decoding Approach to Video Classification | |
| WO2010089488A1 (fr) | Procède de fusion de segments de programmes audiovisuels, dispositif, et produit programme d'ordinateur correspondant | |
| EP1209614A1 (fr) | Procédés de partition d'un ensemble d'objets et procédé de recherche dans une partition d'un ensemble d'objets | |
| WO2017093467A1 (fr) | Procede de gestion de contenus video pour leur edition selectionnant des moments ponctuels et utilisant des modeles adaptifs automatisables | |
| WO2013014290A1 (fr) | Procédé de codage d'une image après redimensionnement par suppression de pixels |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| 17P | Request for examination filed |
Effective date: 20110601 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO SE SI SK SM TR |
|
| RIN1 | Information on inventor provided before grant (corrected) |
Inventor name: BERRANI, SID, AHMED Inventor name: BESSET, BENOIT Inventor name: LECHAT, PATRICK |
|
| RIN1 | Information on inventor provided before grant (corrected) |
Inventor name: LECHAT, PATRICK Inventor name: BESSET, BENOIT Inventor name: BERRANI, SID, AHMED |
|
| DAX | Request for extension of the european patent (deleted) | ||
| RAP1 | Party data changed (applicant data changed or rights of an application transferred) |
Owner name: ORANGE |
|
| 17Q | First examination report despatched |
Effective date: 20140328 |
|
| REG | Reference to a national code |
Ref country code: DE Ref legal event code: R003 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION HAS BEEN REFUSED |
|
| 18R | Application refused |
Effective date: 20180405 |