CN111797912A - System and method for identifying film generation type and construction method of identification model - Google Patents
System and method for identifying film generation type and construction method of identification model Download PDFInfo
- Publication number
- CN111797912A CN111797912A CN202010580262.5A CN202010580262A CN111797912A CN 111797912 A CN111797912 A CN 111797912A CN 202010580262 A CN202010580262 A CN 202010580262A CN 111797912 A CN111797912 A CN 111797912A
- Authority
- CN
- China
- Prior art keywords
- type
- film
- identifying
- pictures
- neural network
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/32—Normalisation of the pattern dimensions
-
- 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
- G06V10/56—Extraction of image or video features relating to colour
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Multimedia (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a system and a method for identifying the type of a film generation and a construction method of an identification model. A system for chronological type identification of a movie, comprising: a trained film generation type identification model; the computing device, the storage device and the input-output device constitute a computer device for automatically identifying the process; a convolution neural network system formed by digital film decoding, screenshot and picture preprocessing; and the classifier can analyze the film generation type according to the model reasoning result. By the system and the method, the generation type of the film can be automatically identified after the computing device obtains the digital film from the storage device. Compared with manual identification, identification by using a computing device has the advantages of reliability, rapidness, low cost, batch processing and the like.
Description
Technical Field
The invention relates to a system and a method for identifying the year type of a film and a construction method of an identification model, belonging to the technical field of information.
Background
In recent years, video services have been rapidly developed, and activities such as tracing a network play and watching digital television have become important entertainment modes for people. Video service providers often store a large number of films, how to automatically, quickly, reliably and batch label the films with the labels of the year, subject, genre and the like is of great significance for helping the service providers to provide high-quality video services.
Disclosure of Invention
The invention aims to provide a system and a method for identifying the type of a film generation and a construction method of an identification model.
In order to achieve the purpose, the invention is realized by the following technical scheme:
a system for chronological type identification of a movie, comprising:
a trained film generation type identification model;
the computing device, the storage device and the input-output device constitute a computer device for automatically identifying the process;
a convolution neural network system formed by digital film decoding, screenshot and picture preprocessing;
and the classifier can analyze the film generation type according to the model reasoning result.
On the basis of the system for identifying the film generation types, the convolutional neural network system is of a VGG-16 network structure.
A method for identifying the generation type of a film by the system comprises the following steps:
reading a movie to be identified from a storage device by the computing device, and cutting off M pictures at uniform time intervals after cutting off the first a minutes and the last b minutes of the movie;
after preprocessing the M pictures, respectively inputting the M pictures into a convolutional neural network;
for each picture input, the convolutional neural network can deduce an N-dimensional probability vector, and each dimension corresponds to the probability that the picture belongs to the corresponding age type;
after all M pictures are reasoned, the system for identifying the film generation type inputs M probability vectors into the classifier.
A method for constructing the film generation type identification model is characterized by comprising the following steps:
s1, constructing a training set and a verification set: preparing a large number of digital films belonging to N types of years, and uniformly intercepting pictures from each film time interval as training data and verification data of the type;
s2 construction of the convolutional neural network: using a deep convolutional network for image classification tasks as the subject of the network;
s3 training of the convolutional neural network: freezing parameters of a convolution base, training a classifier, and using Dropout at the first layer of the classifier; after multiple rounds of training, the classifier achieves better accuracy, the bottom layer of the convolution base is unfrozen, fine tuning training is carried out, and the accuracy of the verification network on the verification set is verified.
According to the construction method of the film generation type identification model, all training pictures and verification pictures are required to be preprocessed, and the method comprises the following steps: 1) scaling to the input size required by the neural network; 2) subtracting the average RGB value of the whole picture data set from the pixel of each picture; 3) the pixel values of RGB are divided by 255 so that the RGB values are between 0 and 1.
6. The method for constructing a model for identifying a chronological genre of a movie as claimed in claim 3, wherein: the classifier analyzes the chronological type of the film according to the following algorithm: any playing time t of the video stream corresponds to one frameTaking the frame as the input of the model can obtain an N-dimensional probability vectorEach dimension of the vector corresponds to a year type respectively;
Integrating the scoring functions of all playing moments to obtain total scoreThe total score is an N-dimensional vector whose dimensions are defined byAnd each dimension corresponds to a year type;
and taking the age type with the maximum total score as the classification result of the video.
The invention has the advantages that: by the system and the method, the generation type of the film can be automatically identified after the computing device obtains the digital film from the storage device, and compared with manual identification, the identification by using the computing device has the advantages of reliability, rapidness, low cost, batch processing and the like.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention.
Fig. 1 is a schematic flow chart of a method for constructing a film generation type identification model according to the present invention.
Fig. 2 is a flowchart illustrating a film generation type identification method according to the present invention.
Fig. 3 is a schematic diagram of the connection of the film generation type identification system according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A system of chronological type identification, comprising: a trained film generation type identification model; the computing device, the storage device and the input-output device constitute a computer device for automatically identifying the process; a convolution neural network system formed by digital film decoding, screenshot and picture preprocessing; and the classifier can analyze the film generation type according to the model reasoning result.
A kind of said system carries on the identifying method of the film year type, this method will wait to discern the film and print year type label, wait to discern the actual year type of the film must be one in N year types preserved, these year types can be types such as ancient clothes, republic of China, contemporary, the method includes the following steps:
reading a movie to be identified from a storage device by the computing device, and cutting off M pictures at uniform time intervals after cutting off the first a minutes and the last b minutes of the movie;
after preprocessing the M pictures, respectively inputting the M pictures into a convolutional neural network;
picture pre-processing includes, but is not limited to, subtracting the average RGB value of all pixels from the RGB value of each pixel of the picture, multiplying 1/255 the RGB values of all pixels so that the values fall within the interval of 0-1, scaling the picture to some fixed size, etc.;
for each picture input, the convolutional neural network can deduce an N-dimensional probability vector, and each dimension corresponds to the probability that the picture belongs to the corresponding age type;
after all M pictures are reasoned, the system for identifying the film generation type inputs M probability vectors into the classifier.
A method for constructing the film generation type identification model comprises the following steps:
s1, constructing a training set and a verification set: for each type of era, S digital movies are prepared, S being large enough; for each movie, the first a minutes and the last b minutes are firstly pinched off to ensure that the leader and the trailer are pinched off; intercepting T pictures at equal time intervals to serve as data of the type; thus, S multiplied by T pictures can be obtained from each type and are distributed to a training set and a verification set according to a certain proportion;
s2 construction of the convolutional neural network: using a deep convolutional network for image classification task as the main body of the network, for example, a VGG16 network, using parameters of a VGG16 network trained on a large picture data set, initializing its own network, for example, using a Keras download trained VGG16 network;
s3 training of the convolutional neural network: freezing parameters of a convolution base, training a classifier, wherein Dropout is used in the first layer of the classifier to reduce the overfitting problem; after multiple rounds of training, the classifier achieves better accuracy, the bottom layer of the convolution base is unfrozen, fine tuning training is carried out, and the accuracy of the verification network on the verification set is verified.
All training pictures and verification pictures are required to be preprocessed, and the method comprises the following steps: 1) scaling to the input size required by the neural network; 2) subtracting the average RGB value of the whole picture data set from the pixel of each picture; 3) the pixel values of RGB are divided by 255 so that the RGB values are between 0 and 1.
The classifier analyzes the chronological type of the film according to the following algorithm: introducing N-dimensional probability vectorsFunction of (2)WhereinHereinafter, will be describedReferred to as a scoring function;
for a video stream S with a duration T, each playing time T corresponds to a frameInputting F into the network to obtain a probability vectorThe scoring function is a unitary function of the probability vector, and the scoring function is also a unitary function during playing according to the chain ruleA univariate function R = R (t) at moment t;
integrating the scores of all the moments of the video stream S to obtain a total scoreThe function is an N-dimensional vector, an
As a result of this, the number of the,i.e. film type, whereinAnd calculating the subscript of the vector.
For digital video, the playing time is discrete, in other words, a digital film with a frame rate of 25fps contains 25 frames per second, rather than an infinite number of frames. In this case, the integral expression is degenerated into a summation expression, i.e.
T on the left of the equal sign is a constant, which has no influence on the classification result, and for convenience, the expression of the total scoring function can be redefined as
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (6)
1. A system for identifying a chronological type of a movie, comprising:
a trained film generation type identification model;
the computing device, the storage device and the input-output device constitute a computer device for automatically identifying the process;
a convolution neural network system formed by digital film decoding, screenshot and picture preprocessing;
and the classifier can analyze the film generation type according to the model reasoning result.
2. The system for dating type of film as claimed in claim 1, wherein: the convolutional neural network system is of a VGG-16 network structure.
3. A method for identifying the chronological type of a film according to the system of claim 1 or 2, comprising the steps of:
reading a movie to be identified from a storage device by the computing device, and cutting off M pictures at uniform time intervals after cutting off the first a minutes and the last b minutes of the movie;
after preprocessing the M pictures, respectively inputting the M pictures into a convolutional neural network;
for each picture input, the convolutional neural network can deduce an N-dimensional probability vector, and each dimension corresponds to the probability that the picture belongs to the corresponding age type;
after all M pictures are reasoned, the system for identifying the film generation type inputs M probability vectors into the classifier.
4. A method for constructing a model for identifying an era type of a movie as claimed in claim 1 or 2, comprising the steps of:
s1, constructing a training set and a verification set: preparing a large number of digital films belonging to N types of years, and uniformly intercepting pictures from each film time interval as training data and verification data of the type;
s2 construction of the convolutional neural network: using a deep convolutional network for image classification tasks as the subject of the network;
s3 training of the convolutional neural network: freezing parameters of a convolution base, training a classifier, and using Dropout at the first layer of the classifier; after multiple rounds of training, the classifier achieves better accuracy, the bottom layer of the convolution base is unfrozen, fine tuning training is carried out, and the accuracy of the verification network on the verification set is verified.
5. The method for constructing a model for identifying a chronological genre of a movie as claimed in claim 3, wherein: all training pictures and verification pictures are required to be preprocessed, and the method comprises the following steps: 1) scaling to the input size required by the neural network; 2) subtracting the average RGB value of the whole picture data set from the pixel of each picture; 3) the pixel values of RGB are divided by 255 so that the RGB values are between 0 and 1.
6. The method for constructing a model for identifying a chronological genre of a movie as claimed in claim 3, wherein: the classifier analyzes the chronological type of the film according to the following algorithm:
any playing time t of the video stream corresponds to one frameTaking the frame as the input of the model can obtain an N-dimensional probability vectorEach dimension of the vector corresponds to a year type respectively;
Integrating the scoring functions of all playing moments to obtain total scoreThe total score is an N-dimensional vector whose dimensions are defined byAnd each dimension corresponds to a year type;
and taking the age type with the maximum total score as the classification result of the video.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010580262.5A CN111797912B (en) | 2020-06-23 | 2020-06-23 | System and method for identifying film age type and construction method of identification model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010580262.5A CN111797912B (en) | 2020-06-23 | 2020-06-23 | System and method for identifying film age type and construction method of identification model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111797912A true CN111797912A (en) | 2020-10-20 |
CN111797912B CN111797912B (en) | 2023-09-22 |
Family
ID=72803677
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010580262.5A Active CN111797912B (en) | 2020-06-23 | 2020-06-23 | System and method for identifying film age type and construction method of identification model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111797912B (en) |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103229223A (en) * | 2010-09-28 | 2013-07-31 | 国际商业机器公司 | Providing answers to questions using multiple models to score candidate answers |
CN105844251A (en) * | 2016-03-31 | 2016-08-10 | 乐视控股(北京)有限公司 | Cartoon video identification method and device |
US20170053024A1 (en) * | 2014-04-28 | 2017-02-23 | Hewlett Packard Enterprise Development Lp | Term chain clustering |
CN107067020A (en) * | 2016-12-30 | 2017-08-18 | 腾讯科技(上海)有限公司 | Image identification method and device |
CN107194419A (en) * | 2017-05-10 | 2017-09-22 | 百度在线网络技术(北京)有限公司 | Video classification methods and device, computer equipment and computer-readable recording medium |
US20180032846A1 (en) * | 2016-08-01 | 2018-02-01 | Nvidia Corporation | Fusing multilayer and multimodal deep neural networks for video classification |
CN108154134A (en) * | 2018-01-11 | 2018-06-12 | 天格科技(杭州)有限公司 | Internet live streaming pornographic image detection method based on depth convolutional neural networks |
CN109753984A (en) * | 2017-11-07 | 2019-05-14 | 北京京东尚科信息技术有限公司 | Video classification methods, device and computer readable storage medium |
CN109919031A (en) * | 2019-01-31 | 2019-06-21 | 厦门大学 | A kind of Human bodys' response method based on deep neural network |
CN110147700A (en) * | 2018-05-18 | 2019-08-20 | 腾讯科技(深圳)有限公司 | Video classification methods, device, storage medium and equipment |
CN110766096A (en) * | 2019-10-31 | 2020-02-07 | 北京金山云网络技术有限公司 | Video classification method and device and electronic equipment |
CN110866593A (en) * | 2019-11-05 | 2020-03-06 | 西南交通大学 | Highway severe weather identification method based on artificial intelligence |
CN110991246A (en) * | 2019-10-31 | 2020-04-10 | 天津市国瑞数码安全系统股份有限公司 | Video detection method and system |
-
2020
- 2020-06-23 CN CN202010580262.5A patent/CN111797912B/en active Active
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103229223A (en) * | 2010-09-28 | 2013-07-31 | 国际商业机器公司 | Providing answers to questions using multiple models to score candidate answers |
US20170053024A1 (en) * | 2014-04-28 | 2017-02-23 | Hewlett Packard Enterprise Development Lp | Term chain clustering |
CN105844251A (en) * | 2016-03-31 | 2016-08-10 | 乐视控股(北京)有限公司 | Cartoon video identification method and device |
US20180032846A1 (en) * | 2016-08-01 | 2018-02-01 | Nvidia Corporation | Fusing multilayer and multimodal deep neural networks for video classification |
CN107067020A (en) * | 2016-12-30 | 2017-08-18 | 腾讯科技(上海)有限公司 | Image identification method and device |
CN107194419A (en) * | 2017-05-10 | 2017-09-22 | 百度在线网络技术(北京)有限公司 | Video classification methods and device, computer equipment and computer-readable recording medium |
CN109753984A (en) * | 2017-11-07 | 2019-05-14 | 北京京东尚科信息技术有限公司 | Video classification methods, device and computer readable storage medium |
CN108154134A (en) * | 2018-01-11 | 2018-06-12 | 天格科技(杭州)有限公司 | Internet live streaming pornographic image detection method based on depth convolutional neural networks |
CN110147700A (en) * | 2018-05-18 | 2019-08-20 | 腾讯科技(深圳)有限公司 | Video classification methods, device, storage medium and equipment |
CN109919031A (en) * | 2019-01-31 | 2019-06-21 | 厦门大学 | A kind of Human bodys' response method based on deep neural network |
CN110766096A (en) * | 2019-10-31 | 2020-02-07 | 北京金山云网络技术有限公司 | Video classification method and device and electronic equipment |
CN110991246A (en) * | 2019-10-31 | 2020-04-10 | 天津市国瑞数码安全系统股份有限公司 | Video detection method and system |
CN110866593A (en) * | 2019-11-05 | 2020-03-06 | 西南交通大学 | Highway severe weather identification method based on artificial intelligence |
Non-Patent Citations (4)
Title |
---|
SICHENG ZHAO等: "Video classification and recommendation based on affective analysis of viewers", 《NEUROCOMPUTING》, vol. 119, pages 101, XP028688389, DOI: 10.1016/j.neucom.2012.04.042 * |
周瑞环: "基于视频流行度及视频分类标签的Top-N推荐", 《中国优秀硕士学位论文全文数据库 信息科技辑》, no. 2018, pages 138 - 1031 * |
朱宁: "基于智能机顶盒视频点播系统的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》, no. 2015, pages 136 - 439 * |
潘翔: "基于多分类器集成的模式识别研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》, no. 2002, pages 138 - 526 * |
Also Published As
Publication number | Publication date |
---|---|
CN111797912B (en) | 2023-09-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11182620B2 (en) | Method for training a convolutional recurrent neural network and for semantic segmentation of inputted video using the trained convolutional recurrent neural network | |
CN111950723B (en) | Neural network model training method, image processing method, device and terminal equipment | |
CN111738357B (en) | Junk picture identification method, device and equipment | |
WO2019147687A1 (en) | Computer vision systems and methods for unsupervised representation learning by sorting sequences | |
US10524005B2 (en) | Facilitating television based interaction with social networking tools | |
US20230353828A1 (en) | Model-based data processing method and apparatus | |
CN111985281B (en) | Image generation model generation method and device and image generation method and device | |
US20220172476A1 (en) | Video similarity detection method, apparatus, and device | |
Su et al. | Image inpainting for random areas using dense context features | |
US20210225005A1 (en) | Selection of Video Frames Using a Machine Learning Predictor | |
Barney et al. | Predicting genre from movie posters | |
Sultana et al. | Unsupervised adversarial learning for dynamic background modeling | |
CN112291634B (en) | Video processing method and device | |
Wang et al. | Archive film defect detection and removal: an automatic restoration framework | |
Liu et al. | Reference-guided texture and structure inference for image inpainting | |
US11526972B2 (en) | Simultaneously correcting image degradations of multiple types in an image of a face | |
Qiu et al. | DiffBFR: Bootstrapping Diffusion Model for Blind Face Restoration | |
CN111444822B (en) | Object recognition method and device, storage medium and electronic device | |
CN116261009B (en) | Video detection method, device, equipment and medium for intelligently converting video audience | |
CN111797912A (en) | System and method for identifying film generation type and construction method of identification model | |
US20210224571A1 (en) | Automated Cropping of Images Using a Machine Learning Predictor | |
CN112488072A (en) | Method, system and equipment for acquiring face sample set | |
CN110826545A (en) | Video category identification method and related device | |
WO2023024424A1 (en) | Segmentation network training method, using method, apparatus, device, and storage medium | |
CN116012550A (en) | Face deformation target correction method and device, equipment, medium and product thereof |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: 250100 West District, North First Floor, S06 Building, 1036 Langchao Road, Jinan High-tech Zone, Shandong Province Applicant after: Shandong Inspur Ultra HD Video Industry Co.,Ltd. Address before: 250101 west area, North 1st floor, S06 building, 1036 Chaochao Road, high tech Zone, Jinan City, Shandong Province Applicant before: SHANDONG YUNMAN INTELLIGENT TECHNOLOGY Co.,Ltd. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |