CN105578186A - Code stream detection method and system for zoom-in scene - Google Patents

Code stream detection method and system for zoom-in scene Download PDF

Info

Publication number
CN105578186A
CN105578186A CN201511026237.8A CN201511026237A CN105578186A CN 105578186 A CN105578186 A CN 105578186A CN 201511026237 A CN201511026237 A CN 201511026237A CN 105578186 A CN105578186 A CN 105578186A
Authority
CN
China
Prior art keywords
frame
scene
camera lens
represent
region
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.)
Pending
Application number
CN201511026237.8A
Other languages
Chinese (zh)
Inventor
舒倩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
SHENZHEN YUNZHOU MULTIMEDIA TECHNOLOGY Co Ltd
Original Assignee
SHENZHEN YUNZHOU MULTIMEDIA TECHNOLOGY Co Ltd
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 SHENZHEN YUNZHOU MULTIMEDIA TECHNOLOGY Co Ltd filed Critical SHENZHEN YUNZHOU MULTIMEDIA TECHNOLOGY Co Ltd
Priority to CN201511026237.8A priority Critical patent/CN105578186A/en
Publication of CN105578186A publication Critical patent/CN105578186A/en
Pending legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/142—Detection of scene cut or scene change
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/14—Picture signal circuitry for video frequency region

Abstract

Embodiments of the invention disclose a code stream detection method and system for a zoom-in scene. According to the method, the zoom-in scene is detected through analyzing code stream information. Through reutilization of code stream compression information, the calculated amount is greatly reduced.

Description

A kind of camera lens promotes code stream detection method and the system of scene
Technical field
The present invention relates to technical field of image processing, particularly relate to code stream detection method and system that a kind of camera lens promotes scene.
Background technology
Decoding end image procossing, transcoding are all the film sources based on compressed bit stream, all containing abundant film source feature in its coded message of carrying.If cast aside these compressed informations, carry out conventional graphical analysis, will the waste of amount of calculation be produced, thus the performance of limit algorithm.On the other hand, the lifting of the performance of decoding end image procossing, transcoder Video coding, also needs a kind of scene detection methods of little amount of calculation stable performance.
Summary of the invention
The object of the embodiment of the present invention is to propose the code stream detection method that a kind of camera lens promotes scene, is intended to the problem solving amount of calculation waste in prior art scene code stream detection method.
The embodiment of the present invention is achieved in that a kind of camera lens promotes the code stream detection method of scene, said method comprising the steps of:
Steps A: if frame tfor I frame, then enter step e; Otherwise, if note t-1=1, then enter step D, otherwise, then enter step B;
Wherein, frame trepresent present frame, play frame also referred to as t; Frame t-1represent that t-1 is play frame; Note t-1represent frame t-1scene switch identifier, the scene of all frames switches identifier initial value and is 0;
Step B: calculate scene and switch identifier;
Step C: if note t=1, then enter step e; Otherwise, enter step D;
Step D: carry out code stream information analysis to present frame, obtains present frame camera lens and promotes scene identifiers;
Step e: if frame t+countexist, then by frame t+countbe set to present frame, then enter steps A; Otherwise, terminate;
Wherein, frame t+countrepresent that t+count is play frame; Count represents the algorithmic technique cycle, and weight is weight factor.
Another object of the embodiment of the present invention is to propose the code stream detection system that a kind of camera lens promotes scene, and described system comprises:
Present frame classification judging treatmenting module, if for judging frame tfor I frame, then enter t+count and play frame judging treatmenting module; Otherwise, if note t-1=1, then enter present frame camera lens and promote scene identifiers acquisition device, otherwise (note t-1=0), then enter scene and switch identifier calculation device;
Wherein, frame trepresent present frame, play frame also referred to as t; Frame t-1represent that t-1 is play frame; Note t-1represent frame t-1scene switch identifier, the scene of all frames switches identifier initial value and is 0;
Scene switches identifier calculation device, switches identifier for calculating scene;
Scene switches identifier judging treatmenting module, if for judging note t=1, then enter t+count and play frame judging treatmenting module; Otherwise, enter present frame camera lens and promote scene identifiers acquisition device;
Present frame camera lens promotes scene identifiers acquisition device, for carrying out code stream information analysis to present frame, obtaining present frame camera lens and promoting scene identifiers;
Play frame judging treatmenting module for t+count, if for judging frame t+countexist, then by frame t+countbe set to present frame, then enter present frame classification judging treatmenting module; Otherwise, terminate;
Wherein, frame t+countrepresent that t+count is play frame; Count represents the algorithmic technique cycle, and fps≤count≤min (fps*weight, num_I), num_I represents I frame period; Min represents and minimizes; Fps represents film source sampling frame per second, and weight is weight factor, 0.5≤weight≤10.
Beneficial effect of the present invention
The present invention proposes code stream detection method and the system that a kind of camera lens promotes scene.The inventive method, by the analysis to code stream information, detects that camera lens promotes scene.Secondary based on code stream compressed information utilizes, thus reaches declining to a great extent of amount of calculation.
Accompanying drawing explanation
Fig. 1 is the code stream detection method flow chart that a kind of camera lens of the preferred embodiment of the present invention promotes scene;
Fig. 2 is the method flow diagram of Step2 in Fig. 1;
Fig. 3 is the method flow diagram of Step4 in Fig. 1;
Fig. 4 is the method flow diagram of Step41 in Fig. 3;
Fig. 5 is the method flow diagram of Step42 in Fig. 3;
Fig. 6 is the code stream detection system structure chart that a kind of camera lens of the preferred embodiment of the present invention promotes scene;
Fig. 7 is that Fig. 6 Scene switches identifier calculation device detailed structure view;
Fig. 8 is that in Fig. 6, present frame camera lens promotes scene identifiers acquisition device detailed structure view;
Fig. 9 is appropriate area acquisition device detailed structure view in Fig. 8;
Figure 10 is regional movement direction acquisition device detailed structure view in Fig. 8.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention being further elaborated, for convenience of explanation, illustrate only the part relevant to the embodiment of the present invention.Should be appreciated that the specific embodiment that this place is described, only for explaining the present invention, not in order to limit the present invention.
The embodiment of the present invention proposes code stream detection method and the system that a kind of camera lens promotes scene.The inventive method, by the analysis to code stream information, detects that camera lens promotes scene.Secondary based on code stream compressed information utilizes, thus reaches declining to a great extent of amount of calculation.
Embodiment one
Fig. 1 is the code stream detection method flow chart that a kind of camera lens of the preferred embodiment of the present invention promotes scene; Said method comprising the steps of:
Step1: if frame tfor I frame, then enter Step5; Otherwise, if note t-1=1, then enter Step4, otherwise, then enter Step2.
Wherein, frame trepresent present frame, play frame also referred to as t; Frame t-1represent that t-1 is play frame; Note t-1represent frame t-1scene switch identifier, the scene of all frames switches identifier initial value and is 0.
Step2: calculate scene and switch identifier.
Fig. 2 is the method flow diagram of Step2 in Fig. 1;
Step21: the broad sense I encoding block quantity calculating present frame.
Above-mentioned broad sense I encoding block quantity computation method:
Number i=sum (sign (cu k| cu kat least comprise the prediction block of an intra prediction mode) | 1≤k≤K).
Wherein, sum (variable | condition) represent the variable summation that satisfies condition; Number irepresent the broad sense I encoding block number of present frame; Cu krepresent a kth coding unit of present frame, k represents position number; The coding unit number that K presentation video frame comprises.
Step22: if number i>Thres 2* K, then note t=1; Otherwise note t=0.
Wherein, Thres 2be the second decision threshold, usual 0.8≤Thres 2≤ 1; Note trepresent frame tscene switch identifier.
Step3: if note t=1, then enter Step5; Otherwise, enter Step4.
Step4: carry out code stream information analysis to present frame, obtains present frame camera lens and promotes scene identifiers.
Fig. 3 is the method flow diagram of Step4 in Fig. 1;
Step41: obtain four suitable regions.
Fig. 4 is the method flow diagram of Step41 in Fig. 3;
Step411: the upper left corner fringe region arranging present frame is first area, arrange the upper right corner fringe region of present frame is second area, arranges that the lower left corner fringe region of present frame is the 3rd region, to arrange the lower right corner fringe region of present frame be the 4th region, is designated as region respectively i, 1≤i≤4.
Wherein, region irepresent the i-th region, require that can not there be intersection in four regions each other.
Step412: to add up respectively in four regions neither the coding unit number of intra-frame prediction block non-Skip block again.
Count i=sum (sign (cu i, ki| condition i) | 1≤ki≤num i), 1≤i≤4
Wherein, count ito represent in the i-th region neither the coding unit number of intra-frame prediction block non-Skip block again; Cu i, kirepresent kth i the coding unit in the i-th region, ki represents position number;
Condition i: and and cu i, ki∈ region iand pu k,m∈ cu i, ki;
Pu ki, mrepresent cu i, kim prediction block; M represents pu ki, mat cu i, kiin position number; num irepresent region ithe coding unit number comprised.
Step413: if count i<num i* Thres 3, then first adopt method in Step411, reset the i-th corresponding region, and then enter Step412; Otherwise, enter Step42.
Wherein, Thres 3be the 3rd decision threshold, usual 0.5≤Thres 3≤ 1.
Step42: obtain four regional movement directions.
Fig. 5 is the method flow diagram of Step42 in Fig. 3;
Step421: the characteristic kinematic vector calculating each encoding block in four regions.
wherein,
Mvxl i, ki=sum (signxl (mv ki, m) * weight ki, m| condition i),
Mvxr i, ki=sum (signxr (mv ki, m) * weight ki, m| condition i),
Mvyu i, ki=sum (signyu (mv ki, m) * weight ki, m| condition i),
Mvyd i, ki=sum (signyd (mv ki, m) * weight ki, m| condition i),
weight ki,m＝size(pu ki,m)/size(min_pu)，1≤i≤4，1≤ki≤num i。
Wherein, mvx i, ki, mvy i, kirepresent cu inside the i-th region respectively i, kithe x-axis component of characteristic kinematic vector, y-axis component; Mv ki, mfor pu ki, mmotion vector; Weight ki, mrepresent pu ki, mweight factor; Size (pu ki, m), size (min_pu) represents pu respectively k,msize, the minimum predicting unit size of present frame.
Step422: the direction of motion direction calculating four regions i=δ (max (diff i j| 1≤j≤4)), 1≤i≤4.
Wherein, direction irepresent region ithe direction of motion; Diff i jrepresent an i-th region jth temporary variable, 1≤j≤4, diff i j=sum (sign (cu i, ki| area condition j) | 1≤ki≤num i); Area condition 1:mvx i, ki<0 and mvy i, ki<0, area condition 2:mvx i, ki>0 and mvy i, ki<0, area condition 3:mvx i, ki<0 and mvy i, ki>0, area condition 4:mvx i, ki>0 and mvy i, ki>0; Max (variable | condition) represent the variable maximizing that satisfies condition; δ ( variable) represent the superscript obtaining variable.
Step43: the scene determining whether there is camera lens promotion according to the direction in four regions, if there is camera lens to promote scene, is then carefully judged to be that camera lens pushes away far away or camera lens furthers further.
If i.e. direction 1=1 and direction 2=2 and direction 3=3 and direction 4=4, then judge currently to promote scene and furthering as camera lens as camera lens, note_py is set t=1;
Otherwise, if direction 1=4 and direction 2=3 and direction 3=2 and direction 4=1, then judge currently to promote scene and pushing away far as camera lens as camera lens, note_py is set t=-1; Otherwise, then judge that there is not camera lens promotes scene, arranges note_py t=0.
Wherein, note_py tfor camera lens promotes scene identifiers.
Step5: if frame t+countexist, then by frame t+countbe set to present frame, then enter Step1; Otherwise, terminate.
Wherein, frame t+countrepresent that t+count is play frame; Count represents the algorithmic technique cycle, and general fps≤count≤min (fps*weight, num_I), num_I represents I frame period; Min represents and minimizes; Fps represents film source sampling frame per second, and weight is weight factor, general 0.5≤weight≤10.
Embodiment two
Fig. 6 is the code stream detection system structure chart that a kind of camera lens of the preferred embodiment of the present invention promotes scene; Described system comprises:
Present frame classification judging treatmenting module, if for judging frame tfor I frame, then enter t+count and play frame judging treatmenting module; Otherwise, if note t-1=1, then enter present frame camera lens and promote scene identifiers acquisition device, otherwise (note t-1=0), then enter scene and switch identifier calculation device.
Wherein, frame trepresent present frame, play frame also referred to as t; Frame t-1represent that t-1 is play frame; Note t-1represent frame t-1scene switch identifier, the scene of all frames switches identifier initial value and is 0.
Scene switches identifier calculation device, switches identifier for calculating scene;
Scene switches identifier judging treatmenting module, if for judging note t=1, then enter t+count and play frame judging treatmenting module; Otherwise, enter present frame camera lens and promote scene identifiers acquisition device;
Present frame camera lens promotes scene identifiers acquisition device, for carrying out code stream information analysis to present frame, obtaining present frame camera lens and promoting scene identifiers;
Play frame judging treatmenting module for t+count, if for judging frame t+countexist, then by frame t+countbe set to present frame, then enter present frame classification judging treatmenting module; Otherwise, terminate.
Wherein, frame t+countrepresent that t+count is play frame; Count represents the algorithmic technique cycle, and general fps≤count≤min (fps*weight, num_I), num_I represents I frame period; Min represents and minimizes; Fps represents film source sampling frame per second, and weight is weight factor, general 0.5≤weight≤10.
Fig. 7 is that Fig. 6 Scene switches identifier calculation device detailed structure view;
Further, described scene switches identifier calculation device and comprises present frame broad sense I encoding block number calculating section, broad sense I encoding block quantity judging treatmenting module,
Present frame broad sense I encoding block number calculating section, for calculating the broad sense I encoding block quantity of present frame.
Above-mentioned broad sense I encoding block quantity computation method:
Number i=sum (sign (cu k| cu kat least comprise the prediction block of an intra prediction mode) | 1≤k≤K).
Wherein, sum (variable | condition) represent the variable summation that satisfies condition; Number irepresent the broad sense I encoding block number of present frame; Cu krepresent a kth coding unit of present frame, k represents position number; The coding unit number that K presentation video frame comprises.
Broad sense I encoding block quantity judging treatmenting module, if for judging number i>Thres 2* K, then note t=1; Otherwise note t=0;
Wherein, Thres 2be the second decision threshold, usual 0.8≤Thres 2≤ 1; Note trepresent frame tscene switch identifier.
Fig. 8 is that in Fig. 6, present frame camera lens promotes scene identifiers acquisition device detailed structure view;
Further, described present frame camera lens promotion scene identifiers acquisition device comprises appropriate area acquisition device, regional movement direction acquisition device, camera lens promotion scene judging treatmenting module
Appropriate area acquisition device, for obtaining four suitable regions;
Regional movement direction acquisition device, for obtaining four regional movement directions;
Camera lens promotes scene judging treatmenting module, for determining whether there is the scene that camera lens promotes according to the direction in four regions, if there is camera lens to promote scene, is then carefully judged to be that camera lens pushes away far away or camera lens furthers further.
If i.e. direction 1=1 and direction 2=2 and direction 3=3 and direction 4=4, then judge currently to promote scene and furthering as camera lens as camera lens, note_py is set t=1;
Otherwise, if direction 1=4 and direction 2=3 and direction 3=2 and direction 4=1, then judge currently to promote scene and pushing away far as camera lens as camera lens, note_py is set t=-1; Otherwise, then judge that there is not camera lens promotes scene, arranges note_py t=0.
Wherein, note_py tfor camera lens promotes scene identifiers.
Fig. 9 is appropriate area acquisition device detailed structure view in Fig. 8;
Further, described appropriate area acquisition device comprises region and arranges module, non-infra-frame prediction non-Skip block forecast unit number statistical module, non-infra-frame prediction non-Skip block forecast unit number threshold decision processing module,
Region arranges module, being first area for arranging the upper left corner fringe region of present frame, the upper right corner fringe region of present frame being set being second area, arranging that the lower left corner fringe region of present frame is the 3rd region, to arrange the lower right corner fringe region of present frame be the 4th region, be designated as region respectively i, 1≤i≤4.
Wherein, region irepresent the i-th region, require that can not there be intersection in four regions each other.
Non-infra-frame prediction non-Skip block forecast unit number statistical module, for add up respectively in four regions neither the coding unit number of intra-frame prediction block non-Skip block again;
Count i=sum (sign (cu i, ki| condition i) | 1≤ki≤num i), 1≤i≤4
Wherein, count ito represent in the i-th region neither the coding unit number of intra-frame prediction block non-Skip block again; Cu i, kirepresent kth i the coding unit in the i-th region, ki represents position number; Condition i: and and cu i, ki∈ region iand pu k,m∈ cu i, ki; Pu ki, mrepresent cu i, kim prediction block; M represents pu ki, mat cu i, kiin position number; num irepresent region ithe coding unit number comprised.Non-infra-frame prediction non-Skip block forecast unit number threshold decision processing module, if for judging count i<num i* Thres 3, then first adopt region that the processing method of module is set, reset the i-th corresponding region, and then enter non-infra-frame prediction non-Skip block forecast unit number statistical module; Otherwise, enter regional movement direction acquisition device;
Wherein, Thres 3be the 3rd decision threshold, usual 0.5≤Thres 3≤ 1.
Figure 10 is regional movement direction acquisition device detailed structure view in Fig. 8.
Further, described regional movement direction acquisition device comprises characteristic kinematic vectors calculation module, direction of motion computing module,
Characteristic kinematic vectors calculation module, for calculating the characteristic kinematic vector of each encoding block in four regions.
wherein,
Mvxl i, ki=sum (signxl (mv ki, m) * weight ki, m| condition i),
Mvxr i, ki=sum (signxr (mv ki, m) * weight ki, m| condition i),
Mvyu i, ki=sum (signyu (mv ki, m) * weight ki, m| condition i),
Mvyd i, ki=sum (signyd (mv ki, m) * weight ki, m| condition i),
weight ki,m＝size(pu ki,m)/size(min_pu)，1≤i≤4，1≤ki≤num i。
Wherein, mvx i, ki, mvy i, kirepresent cu inside the i-th region respectively i, kithe x-axis component of characteristic kinematic vector, y-axis component; Mv ki, mfor pu ki, mmotion vector; Weight ki, mrepresent pu ki, mweight factor; Size (pu ki, m), size (min_pu) represents pu respectively k,msize, the minimum predicting unit size of present frame.
Direction of motion computing module, for calculating the direction of motion direction in four regions i=δ (max (diff i j| 1≤j≤4)), 1≤i≤4.
Wherein, direction irepresent region ithe direction of motion; Diff i jrepresent an i-th region jth temporary variable, 1≤j≤4, diff i j=sum (sign (cu i, ki| area condition j) | 1≤ki≤num i); Area condition 1:mvx i, ki<0 and mvy i, ki<0, area condition 2:mvx i, ki>0 and mvy i, ki<0, area condition 3:mvx i, ki<0 and mvy i, ki>0, area condition 4:mvx i, ki>0 and mvy i, ki>0; Max (variable | condition) represent the variable maximizing that satisfies condition; δ (variable) represents the superscript obtaining variable.
Those having ordinary skill in the art will appreciate that, the all or part of step realized in above-described embodiment method can have been come by program command related hardware, described program can be stored in a computer read/write memory medium, and described storage medium can be ROM, RAM, disk, CD etc.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.

Claims (11)

1. camera lens promotes a code stream detection method for scene, it is characterized in that, said method comprising the steps of:
Steps A: if frame tfor I frame, then enter step e; Otherwise, if note t-1=1, then enter step D, otherwise, then enter step B;
Wherein, frame trepresent present frame, play frame also referred to as t; Frame t-1represent that t-1 is play frame; Note t-1represent frame t-1scene switch identifier, the scene of all frames switches identifier initial value and is 0;
Step B: calculate scene and switch identifier;
Step C: if note t=1, then enter step e; Otherwise, enter step D;
Step D: carry out code stream information analysis to present frame, obtains present frame camera lens and promotes scene identifiers;
Step e: if frame t+countexist, then by frame t+countbe set to present frame, then enter steps A; Otherwise, terminate;
Wherein, frame t+countrepresent that t+count is play frame; Count represents the algorithmic technique cycle, and weight is weight factor.
2. camera lens as claimed in claim 1 promotes the code stream detection method of scene, it is characterized in that,
Fps≤count≤min (fps*weight, num_I), num_I represents I frame period; Min represents and minimizes; Fps represents film source sampling frame per second, 0.5≤weight≤10.
3. camera lens as claimed in claim 2 promotes the code stream detection method of scene, it is characterized in that,
Described calculating scene switches identifier and is specially:
Calculate the broad sense I encoding block quantity of present frame;
Described broad sense I encoding block quantity computation method:
Number i=sum (sign (cu k| cu kat least comprise the prediction block of an intra prediction mode) | 1≤k≤K);
Wherein, sum (variable | condition) represent the variable summation that satisfies condition; Number irepresent the broad sense I encoding block number of present frame; Cu krepresent a kth coding unit of present frame, k represents position number; The coding unit number that K presentation video frame comprises;
If judge number i>Thres 2* K, then make note t=1; Otherwise make note t=0;
Wherein, Thres 2be the second decision threshold, 0.8≤Thres 2≤ 1; Note trepresent frame tscene switch identifier.
4. camera lens as claimed in claim 3 promotes the code stream detection method of scene, it is characterized in that,
Described code stream information analysis is carried out to present frame, obtains present frame camera lens and promote scene identifiers and be specially:
Obtain four suitable region region i, 1≤i≤4;
Obtain four regional movement direction direction i;
Determine whether there is the scene of camera lens promotion according to the direction in four regions, if there is camera lens to promote scene, be then carefully judged to be that camera lens pushes away far away or camera lens furthers further;
Be specially: if direction 1=1 and direction 2=2 and direction 3=3 and direction 4=4, then judge currently to promote scene and furthering as camera lens as camera lens, note_py is set t=1;
Otherwise, if direction 1=4 and direction 2=3 and direction 3=2 and direction 4=1, then judge currently to promote scene and pushing away far as camera lens as camera lens, note_py is set t=-1; Otherwise, then judge that there is not camera lens promotes scene, arranges note_py t=0;
Wherein, note_py tfor camera lens promotes scene identifiers.
5. camera lens as claimed in claim 4 promotes the code stream detection method of scene, it is characterized in that,
Suitable four regions of described acquisition are specially:
Step411: the upper left corner fringe region arranging present frame is first area, arrange the upper right corner fringe region of present frame is second area, arranges that the lower left corner fringe region of present frame is the 3rd region, to arrange the lower right corner fringe region of present frame be the 4th region, is designated as region respectively i, 1≤i≤4;
Wherein, region irepresent the i-th region, can not there be intersection in four regions each other;
Step412: to add up in four regions neither the coding unit number of intra-frame prediction block non-Skip block again respectively;
Count i=sum (sign (cu i, ki| condition i) | 1≤ki≤num i), 1≤i≤4;
Wherein, count ito represent in the i-th region neither the coding unit number of intra-frame prediction block non-Skip block again; Cu i, kirepresent kth i the coding unit in the i-th region, ki represents position number;
Condition i: and and cu i, ki∈ region iand pu k,m∈ cu i, ki;
Pu ki, mrepresent cu i, kim prediction block; M represents pu ki, mat cu i, kiin position number; num irepresent region ithe coding unit number comprised;
Step413: if count i<num i* Thres 3, then first adopt method in Step411, reset the i-th corresponding region, and then enter Step412; Otherwise, enter step and " obtain four regional movement direction direction i";
Wherein, Thres 3be the 3rd decision threshold, 0.5≤Thres 3≤ 1.
6. camera lens as claimed in claim 5 promotes the code stream detection method of scene, it is characterized in that,
Described acquisition four regional movement directions are specially:
Calculate the characteristic kinematic vector of each encoding block in four regions;
wherein,
Mvxl i, ki=sum (signxl (mv ki, m) * weight ki, m| condition i),
Mvxr i, ki=sum (signxr (mv ki, m) * weight ki, m| condition i),
Mvyu i, ki=sum (signyu (mv ki, m) * weight ki, m| condition i),
Mvyd i, ki=sum (signyd (mv ki, m) * weight ki, m| condition i),
weight ki,m＝size(pu ki,m)/size(min_pu)，1≤i≤4，1≤ki≤num i；
Wherein, mvx i, ki, mvy i, kirepresent cu inside the i-th region respectively i, kithe x-axis component of characteristic kinematic vector, y-axis component; Mv ki, mfor pu ki, mmotion vector; Weight ki, mrepresent pu ki, mweight factor; Size (pu ki, m), size (min_pu) represents pu respectively k,msize, the minimum predicting unit size of present frame;
Calculate the direction of motion direction in four regions i=δ (max (diff i j| 1≤j≤4)), 1≤i≤4;
Wherein, direction irepresent region ithe direction of motion; Diff i jrepresent an i-th region jth temporary variable, 1≤j≤4, diff i j=sum (sign (cu i, ki| area condition j) | 1≤ki≤num i); Area condition 1:mvx i, ki<0 and mvy i, ki<0, area condition 2:mvx i, ki>0 and mvy i, ki<0, area condition 3:mvx i, ki<0 and mvy i, ki>0, area condition 4:mvx i, ki>0 and mvy i, ki>0; Max (variable | condition) represent the variable maximizing that satisfies condition; δ (variable) represents the superscript obtaining variable.
7. camera lens promotes a code stream detection system for scene, and it is characterized in that, described system comprises:
Present frame classification judging treatmenting module, if for judging frame tfor I frame, then enter t+count and play frame judging treatmenting module; Otherwise, if note t-1=1, then enter present frame camera lens and promote scene identifiers acquisition device, otherwise (note t-1=0), then enter scene and switch identifier calculation device;
Wherein, frame trepresent present frame, play frame also referred to as t; Frame t-1represent that t-1 is play frame; Note t-1represent frame t-1scene switch identifier, the scene of all frames switches identifier initial value and is 0;
Scene switches identifier calculation device, switches identifier for calculating scene;
Scene switches identifier judging treatmenting module, if for judging note t=1, then enter t+count and play frame judging treatmenting module; Otherwise, enter present frame camera lens and promote scene identifiers acquisition device;
Present frame camera lens promotes scene identifiers acquisition device, for carrying out code stream information analysis to present frame, obtaining present frame camera lens and promoting scene identifiers;
Play frame judging treatmenting module for t+count, if for judging frame t+countexist, then by frame t+countbe set to present frame, then enter present frame classification judging treatmenting module; Otherwise, terminate;
Wherein, frame t+countrepresent that t+count is play frame; Count represents the algorithmic technique cycle, and fps≤count≤min (fps*weight, num_I), num_I represents I frame period; Min represents and minimizes; Fps represents film source sampling frame per second, and weight is weight factor, 0.5≤weight≤10; .。
8. camera lens as claimed in claim 7 promotes the code stream detection system of scene, it is characterized in that,
Described scene switches identifier calculation device and comprises present frame broad sense I encoding block number calculating section, broad sense I encoding block quantity judging treatmenting module,
Present frame broad sense I encoding block number calculating section, for calculating the broad sense I encoding block quantity of present frame;
Above-mentioned broad sense I encoding block quantity computation method:
Number i=sum (sign (cu k| cu kat least comprise the prediction block of an intra prediction mode) | 1≤k≤K);
Wherein, sum (variable | condition) represent the variable summation that satisfies condition; Number irepresent the broad sense I encoding block number of present frame; Cu krepresent a kth coding unit of present frame, k represents position number; The coding unit number that K presentation video frame comprises;
Broad sense I encoding block quantity judging treatmenting module, if for judging number i>Thres 2* K, then note t=1; Otherwise note t=0;
Wherein, Thres 2be the second decision threshold, 0.8≤Thres 2≤ 1; Note trepresent frame tscene switch identifier.
9. camera lens as claimed in claim 7 promotes the code stream detection system of scene, it is characterized in that,
Described present frame camera lens promotes scene identifiers acquisition device and comprises appropriate area acquisition device, regional movement direction acquisition device, camera lens promotion scene judging treatmenting module
Appropriate area acquisition device, for obtaining four suitable regions;
Regional movement direction acquisition device, for obtaining four regional movement directions;
Camera lens promotes scene judging treatmenting module, for determining whether there is the scene that camera lens promotes according to the direction in four regions, if there is camera lens to promote scene, is then carefully judged to be that camera lens pushes away far away or camera lens furthers further;
Be specially: if direction 1=1 and direction 2=2 and direction 3=3 and direction 4=4, then judge currently to promote scene and furthering as camera lens as camera lens, note_py is set t=1;
Otherwise, if direction 1=4 and direction 2=3 and direction 3=2 and direction 4=1, then judge currently to promote scene and pushing away far as camera lens as camera lens, note_py is set t=-1; Otherwise, then judge that there is not camera lens promotes scene, arranges note_py t=0;
Wherein, note_py tfor camera lens promotes scene identifiers.
10. camera lens as claimed in claim 9 promotes the code stream detection system of scene, it is characterized in that, described appropriate area acquisition device comprises region and arranges module, non-infra-frame prediction non-Skip block forecast unit number statistical module, non-infra-frame prediction non-Skip block forecast unit number threshold decision processing module
Region arranges module, being first area for arranging the upper left corner fringe region of present frame, the upper right corner fringe region of present frame being set being second area, arranging that the lower left corner fringe region of present frame is the 3rd region, to arrange the lower right corner fringe region of present frame be the 4th region, be designated as region respectively i, 1≤i≤4;
Wherein, region irepresent the i-th region, can not there be intersection in four regions each other;
Non-infra-frame prediction non-Skip block forecast unit number statistical module, for add up respectively in four regions neither the coding unit number of intra-frame prediction block non-Skip block again;
Count i=sum (sign (cu i, ki| condition i) | 1≤ki≤num i), 1≤i≤4;
Wherein, count ito represent in the i-th region neither the coding unit number of intra-frame prediction block non-Skip block again; Cu i, kirepresent kth i the coding unit in the i-th region, ki represents position number; Condition i: and and cu i, ki∈ region iand pu k,m∈ cu i, ki; Pu ki, mrepresent cu i, kim prediction block; M represents pu ki, mat cu i, kiin position number; num irepresent region ithe coding unit number comprised; Non-infra-frame prediction non-Skip block forecast unit number threshold decision processing module, if for judging count i<num i* Thres 3, then first adopt region that the processing method of module is set, reset the i-th corresponding region, and then enter non-infra-frame prediction non-Skip block forecast unit number statistical module; Otherwise, enter regional movement direction acquisition device;
Wherein, Thres 3be the 3rd decision threshold, 0.5≤Thres 3≤ 1.
11. camera lenses as claimed in claim 9 promote the code stream detection system of scene, it is characterized in that,
Described regional movement direction acquisition device comprises characteristic kinematic vectors calculation module, direction of motion computing module,
Characteristic kinematic vectors calculation module, for calculating the characteristic kinematic vector of each encoding block in four regions;
wherein,
Mvxl i, ki=sum (signxl (mv ki, m) * weight ki, m| condition i),
Mvxr i, ki=sum (signxr (mv ki, m) * weight ki, m| condition i),
Mvyu i, ki=sum (signyu (mv ki, m) * weight ki, m| condition i),
Mvyd i, ki=sum (signyd (mv ki, m) * weight ki, m| condition i),
weight ki,m＝size(pu ki,m)/size(min_pu)，1≤i≤4，1≤ki≤num i。
Wherein, mvx i, ki, mvy i, kirepresent cu inside the i-th region respectively i, kithe x-axis component of characteristic kinematic vector, y-axis component; Mv ki, mfor pu ki, mmotion vector; Weight ki, mrepresent pu ki, mweight factor; Size (pu ki, m), size (min_pu) represents pu respectively k,msize, the minimum predicting unit size of present frame;
Direction of motion computing module, for calculating the direction of motion direction in four regions i=δ (max (diff i j| 1≤j≤4)), 1≤i≤4;
Wherein, direction irepresent region ithe direction of motion; Diff i jrepresent an i-th region jth temporary variable, 1≤j≤4, diff i j=sum (sign (cu i, ki| area condition j) | 1≤ki≤num i); Area condition 1:mvx i, ki<0 and mvy i, ki<0, area condition 2:mvx i, ki>0 and mvy i, ki<0, area condition 3:mvx i, ki<0 and mvy i, ki>0, area condition 4:mvx i, ki>0 and mvy i, ki>0; Max (variable | condition) represent the variable maximizing that satisfies condition; δ (variable) represents the superscript obtaining variable.
CN201511026237.8A 2015-12-30 2015-12-30 Code stream detection method and system for zoom-in scene Pending CN105578186A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201511026237.8A CN105578186A (en) 2015-12-30 2015-12-30 Code stream detection method and system for zoom-in scene

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201511026237.8A CN105578186A (en) 2015-12-30 2015-12-30 Code stream detection method and system for zoom-in scene

Publications (1)

Publication Number Publication Date
CN105578186A true CN105578186A (en) 2016-05-11

Family

ID=55887783

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201511026237.8A Pending CN105578186A (en) 2015-12-30 2015-12-30 Code stream detection method and system for zoom-in scene

Country Status (1)

Country Link
CN (1) CN105578186A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110636330A (en) * 2019-09-20 2019-12-31 深圳市梦网百科信息技术有限公司 Film source scene complexity analysis method and system based on compressed information
CN112203092A (en) * 2020-09-27 2021-01-08 深圳市梦网视讯有限公司 Code stream analysis method, system and equipment for global motion scene

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH10304371A (en) * 1997-04-25 1998-11-13 Sharp Corp Moving vector detecting device
CN1925617A (en) * 2005-08-29 2007-03-07 三星电子株式会社 Motion estimation method, video encoding method and apparatus using the same
US20090097546A1 (en) * 2007-10-10 2009-04-16 Chang-Hyun Lee System and method for enhanced video communication using real-time scene-change detection for control of moving-picture encoding data rate
CN103227925A (en) * 2013-05-13 2013-07-31 深圳市云宙多媒体技术有限公司 Adaptive frame-B encoding method and system
CN103686184A (en) * 2013-11-18 2014-03-26 深圳市云宙多媒体技术有限公司 Adjusting method and system for frame type in trans-coding
CN104333726A (en) * 2014-10-09 2015-02-04 深圳市云宙多媒体技术有限公司 Deinterlacing method and deinterlacing system for transcoding video source
CN106303528A (en) * 2015-06-04 2017-01-04 北京中传视讯科技有限公司 A kind of code stream translation scene detection method and system

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH10304371A (en) * 1997-04-25 1998-11-13 Sharp Corp Moving vector detecting device
CN1925617A (en) * 2005-08-29 2007-03-07 三星电子株式会社 Motion estimation method, video encoding method and apparatus using the same
US20090097546A1 (en) * 2007-10-10 2009-04-16 Chang-Hyun Lee System and method for enhanced video communication using real-time scene-change detection for control of moving-picture encoding data rate
CN103227925A (en) * 2013-05-13 2013-07-31 深圳市云宙多媒体技术有限公司 Adaptive frame-B encoding method and system
CN103686184A (en) * 2013-11-18 2014-03-26 深圳市云宙多媒体技术有限公司 Adjusting method and system for frame type in trans-coding
CN104333726A (en) * 2014-10-09 2015-02-04 深圳市云宙多媒体技术有限公司 Deinterlacing method and deinterlacing system for transcoding video source
CN106303528A (en) * 2015-06-04 2017-01-04 北京中传视讯科技有限公司 A kind of code stream translation scene detection method and system

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110636330A (en) * 2019-09-20 2019-12-31 深圳市梦网百科信息技术有限公司 Film source scene complexity analysis method and system based on compressed information
CN110636330B (en) * 2019-09-20 2021-10-26 深圳市梦网视讯有限公司 Film source scene complexity analysis method and system based on compressed information
CN112203092A (en) * 2020-09-27 2021-01-08 深圳市梦网视讯有限公司 Code stream analysis method, system and equipment for global motion scene
CN112203092B (en) * 2020-09-27 2024-01-30 深圳市梦网视讯有限公司 Code stream analysis method, system and equipment for global motion scene

Similar Documents

Publication Publication Date Title
US20220353525A1 (en) Image encoding method and apparatus, and image decoding method and apparatus
CN102576411B (en) Scene change detection
CN101409831B (en) Method for processing multimedia video object
Gao et al. The IEEE 1857 standard: Empowering smart video surveillance systems
US8036263B2 (en) Selecting key frames from video frames
CN103024445B (en) The video transcoding method in high in the clouds and Cloud Server
CN104717510A (en) Method and device for processing image
CN103179402A (en) Video compression coding and decoding method and device
TW201406165A (en) Method for decoding image
JP2004207802A (en) Image data processing apparatus and method, recording medium, and program
Zhang et al. Low-complexity and high-efficiency background modeling for surveillance video coding
CN106303528A (en) A kind of code stream translation scene detection method and system
JP2009147911A (en) Video data compression preprocessing method, video data compression method employing the same and video data compression system
Zhang et al. Quadtree decision for depth intra coding in 3D-HEVC by good feature
CN102215387B (en) Video image processing method and coder/decoder
JP2012239085A (en) Image processor, and image processing method
CN105578186A (en) Code stream detection method and system for zoom-in scene
Yuan et al. AccDecoder: Accelerated decoding for neural-enhanced video analytics
CN103327339B (en) The coded method of a kind of intra-frame prediction block division and system
CN103517077A (en) Method and device for rapidly selecting prediction mode
US8897368B2 (en) Image coding device, image coding method, image coding integrated circuit and image coding program
CN103002284A (en) Video coding and decoding method based self-adaption updating of scene model
KR102557904B1 (en) The Method of Detecting Section in which a Movement Frame Exists
CN104333726A (en) Deinterlacing method and deinterlacing system for transcoding video source
CN110062243B (en) Light field video motion estimation method based on neighbor optimization

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20160511