CN106874418A - A kind of automobile model data set method for building up for serving deep learning - Google Patents
A kind of automobile model data set method for building up for serving deep learning Download PDFInfo
- Publication number
- CN106874418A CN106874418A CN201710053285.9A CN201710053285A CN106874418A CN 106874418 A CN106874418 A CN 106874418A CN 201710053285 A CN201710053285 A CN 201710053285A CN 106874418 A CN106874418 A CN 106874418A
- Authority
- CN
- China
- Prior art keywords
- automobile
- file
- picture
- building
- model
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/51—Indexing; Data structures therefor; Storage structures
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Abstract
The present invention provides a kind of automobile model data set method for building up for serving deep learning, comprises the following steps:S1, the multi-angle picture for collecting various car models, every kind of automobile model at least seven classes difference angle picture;S2, tree structure is set up according to the sorting technique in " automobile brand automotive type automobile model automobile time " to picture;S3, the picture under each automobile time is divided into two parts, and training set file is set up using a part, test set file is set up using another part;S4, structure training set file label file and test set file label file, set up data set.The present invention sets up tree structure to auto graph according to the method in " automobile brand automotive type automobile model automobile time ", is favorably improved recognition capability of the system to vehicle, significantly improves the precision of identification;Notebook data integrates multi-angle picture of the picture as automobile, contributes to sophisticated model, improves discrimination.
Description
Technical field
The present invention relates to data classification technology field, and in particular to a kind of automobile model data set for serving deep learning
Method for building up.
Background technology
The recognition accuracy of existing model recognition system is not universal high, and sorting technique is improper when mainly being set up by data set draws
Rise, such as famous vehicle database CompCars, its data structure is " automobile brand-automotive type-year of manufacture ", the number
Cause recognition correct rate not high according to structure.
Therefore, a kind of correct method for effectively building car data collection is needed badly, so as to improve recognition accuracy.
The content of the invention
It is an object of the invention to provide a kind of automobile model data set method for building up for serving deep learning, the method
Problem above can well be solved.
To reach above-mentioned requirements, the present invention is adopted the technical scheme that:A kind of automobile type for serving deep learning is provided
Number collection method for building up, comprises the following steps:
S1, the multi-angle picture for collecting various car models;
S2, set up tree-like knot according to the sorting technique of " automobile brand-automotive type-automobile model-automobile time " to stating picture
Structure;
S3, the picture under each automobile time is divided into two parts, and training set file is set up using a part, use another portion
Divide and set up test set file;
S4, structure training set file label file and test set file label file, set up data set.
Compared with prior art, present invention has the advantage that:
(1)Auto graph is set up according to the level Four sorting technique of " automobile brand-automotive type-automobile model-automobile time "
Tree structure, is favorably improved recognition capability of the system to vehicle, significantly improves the precision of identification;
(2)Notebook data integrates multi-angle picture of the picture as automobile, contributes to sophisticated model, improves discrimination.
Brief description of the drawings
Accompanying drawing described herein is used for providing further understanding of the present application, the part of the application is constituted, at this
Same or analogous part, the schematic description and description of the application are represented using identical reference number in a little accompanying drawings
For explaining the application, the improper restriction to the application is not constituted.In the accompanying drawings:
Fig. 1 is flow chart of the invention.
Specific embodiment
To make the purpose, technical scheme and advantage of the application clearer, below in conjunction with drawings and the specific embodiments, to this
Application is described in further detail.For the sake of simplicity, eliminated well known to a person skilled in the art some skills in below describing
Art feature.
A kind of automobile model data set method for building up for serving deep learning is provided, is comprised the following steps:
S1, the multi-angle picture for collecting various car models, picture can be obtained, or voluntarily to automobile for online collection
Take pictures and obtain;Every kind of car model at least has 7 kinds of pictures of different angles, including following several angles but is not limited to, preceding, left
Before, right, side, after, it is left back, behind the right side etc.;
S2, to above-mentioned multi-angle picture according to " automobile brand-automotive type-automobile model-automobile time " level Four classification side
Method sets up tree structure, for the ease of difference, specification arrangement is carried out in units of file;Specific method is:
Some automobile brand files are set up, with automobile brand as file designation;
Some automotive type files are set up under automobile brand file, with automotive type as file designation;
Some automobile model files are set up under automotive type file, is named by file of automobile model;
Some automobile time files are set up under automobile model file, is named by file of the automobile time;
The picture that step S1 is collected is put in corresponding automobile time file respectively.
S3, the picture in each automobile time file is randomly divided into two parts, and training set text is set up using a part
Part, test set file is set up using another part, and general training collection file, picture quantity should be the 4 of test set file, picture quantity
To 5 times;Specifically include following steps:
Set up training set file;
Some files suitable with automobile time quantity are set up under training set file;
In the most of picture reproduction in each the automobile time file in step S2 to training set file;
Set up test set file;
By remaining picture reproduction in each automobile time file to test set file.
S4, structure training set file label file and test set file label file, set up data set;Specifically include:
Label file is set up under training set file(label), for calling picture in training set;
Label file is set up under test set file(label), for calling picture in test set.
Embodiment described above only represents several embodiments of the invention, and its description is more specific and detailed, but not
It is understood that to be limitation of the scope of the invention.It should be pointed out that for the person of ordinary skill of the art, not departing from
On the premise of present inventive concept, various modifications and improvements can be made, these belong to the scope of the present invention.Therefore this hair
Bright protection domain should be defined by the claim.
Claims (3)
1. a kind of automobile model data set method for building up for serving deep learning, it is characterised in that comprise the following steps:
S1, the multi-angle picture for collecting various car models;
S2, tree-like is set up according to the sorting technique of " automobile brand-automotive type-automobile model-automobile time " to the picture
Structure;
S3, the picture under each automobile time is divided into two parts, and training set file is set up using a part, use another portion
Divide and set up test set file;
S4, structure training set file label file and test set file label file, set up data set.
2. the automobile model data set method for building up for serving deep learning according to claim 1, it is characterised in that institute
Every kind of car model at least has 7 kinds of pictures of different angles in stating step S1.
3. the automobile model data set method for building up for serving deep learning according to claim 1, it is characterised in that institute
State 4-5 times that training set file, picture quantity is test set file, picture quantity.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710053285.9A CN106874418A (en) | 2017-01-24 | 2017-01-24 | A kind of automobile model data set method for building up for serving deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710053285.9A CN106874418A (en) | 2017-01-24 | 2017-01-24 | A kind of automobile model data set method for building up for serving deep learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106874418A true CN106874418A (en) | 2017-06-20 |
Family
ID=59158864
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710053285.9A Pending CN106874418A (en) | 2017-01-24 | 2017-01-24 | A kind of automobile model data set method for building up for serving deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106874418A (en) |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103198164A (en) * | 2013-04-28 | 2013-07-10 | 深圳市元征科技股份有限公司 | Automobile fault file server, management system and management method |
CN103370682A (en) * | 2011-01-04 | 2013-10-23 | Ppg工业俄亥俄公司 | WEB-based color selection system |
US20140109080A1 (en) * | 2012-03-14 | 2014-04-17 | Flextronics Ap, Llc | Self-configuring vehicle console application store |
CN104488004A (en) * | 2012-05-23 | 2015-04-01 | 实耐宝公司 | Methods and systems for providing vehicle repair information |
CN104573184A (en) * | 2014-12-10 | 2015-04-29 | 南车青岛四方机车车辆股份有限公司 | Construction method and construction device for high speed train product meta-model |
CN104751190A (en) * | 2015-04-23 | 2015-07-01 | 武汉大学 | Vehicle part positioning method for vehicle fine recognition |
CN105335702A (en) * | 2015-10-15 | 2016-02-17 | 浙江捷尚视觉科技股份有限公司 | Bayonet vehicle type recognition method based on statistical learning |
CN105574543A (en) * | 2015-12-16 | 2016-05-11 | 武汉烽火众智数字技术有限责任公司 | Vehicle brand and model identifying method and system based on deep learning |
CN105740906A (en) * | 2016-01-29 | 2016-07-06 | 中国科学院重庆绿色智能技术研究院 | Depth learning based vehicle multi-attribute federation analysis method |
CN105894819A (en) * | 2016-01-26 | 2016-08-24 | 浙江捷尚视觉科技股份有限公司 | Fake-license-plate vehicle identification method based on twice verification |
CN105975941A (en) * | 2016-05-31 | 2016-09-28 | 电子科技大学 | Multidirectional vehicle model detection recognition system based on deep learning |
CN106096144A (en) * | 2016-06-13 | 2016-11-09 | 大连理工大学 | A kind of automobile brand gene analysis method based on front face moulding |
-
2017
- 2017-01-24 CN CN201710053285.9A patent/CN106874418A/en active Pending
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103370682A (en) * | 2011-01-04 | 2013-10-23 | Ppg工业俄亥俄公司 | WEB-based color selection system |
US20140109080A1 (en) * | 2012-03-14 | 2014-04-17 | Flextronics Ap, Llc | Self-configuring vehicle console application store |
CN104488004A (en) * | 2012-05-23 | 2015-04-01 | 实耐宝公司 | Methods and systems for providing vehicle repair information |
CN103198164A (en) * | 2013-04-28 | 2013-07-10 | 深圳市元征科技股份有限公司 | Automobile fault file server, management system and management method |
CN104573184A (en) * | 2014-12-10 | 2015-04-29 | 南车青岛四方机车车辆股份有限公司 | Construction method and construction device for high speed train product meta-model |
CN104751190A (en) * | 2015-04-23 | 2015-07-01 | 武汉大学 | Vehicle part positioning method for vehicle fine recognition |
CN105335702A (en) * | 2015-10-15 | 2016-02-17 | 浙江捷尚视觉科技股份有限公司 | Bayonet vehicle type recognition method based on statistical learning |
CN105574543A (en) * | 2015-12-16 | 2016-05-11 | 武汉烽火众智数字技术有限责任公司 | Vehicle brand and model identifying method and system based on deep learning |
CN105894819A (en) * | 2016-01-26 | 2016-08-24 | 浙江捷尚视觉科技股份有限公司 | Fake-license-plate vehicle identification method based on twice verification |
CN105740906A (en) * | 2016-01-29 | 2016-07-06 | 中国科学院重庆绿色智能技术研究院 | Depth learning based vehicle multi-attribute federation analysis method |
CN105975941A (en) * | 2016-05-31 | 2016-09-28 | 电子科技大学 | Multidirectional vehicle model detection recognition system based on deep learning |
CN106096144A (en) * | 2016-06-13 | 2016-11-09 | 大连理工大学 | A kind of automobile brand gene analysis method based on front face moulding |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104536953B (en) | A kind of recognition methods of text emotional valence and device | |
CN105788264B (en) | The method and system of bayonet socket data area crash analysis based on Hbase | |
CN107766787A (en) | Face character recognition methods, device, terminal and storage medium | |
CN105787025B (en) | Network platform public account classification method and device | |
CN106250858A (en) | A kind of recognition methods merging multiple face recognition algorithms and system | |
CN105095884B (en) | A kind of pedestrian's identifying system and processing method based on random forest support vector machines | |
CN110517500B (en) | Man-vehicle association processing method and device | |
CN108021908B (en) | Face age group identification method and device, computer device and readable storage medium | |
CN101980210A (en) | Marked word classifying and grading method and system | |
CN106205120B (en) | A kind of method and device for extracting road cross traffic limitation | |
CN106203296B (en) | The video actions recognition methods of one attribute auxiliary | |
CN105095415A (en) | Method and apparatus for confirming network emotion | |
CN104462550B (en) | Pedestrian's recognition methods again of sorting consistence is merged based on similitude and dissimilarity | |
CN104615986A (en) | Method for utilizing multiple detectors to conduct pedestrian detection on video images of scene change | |
CN106649696A (en) | Information classification method and device | |
CN108959265A (en) | Cross-domain texts sensibility classification method, device, computer equipment and storage medium | |
CN106506528A (en) | A kind of Network Safety Analysis system under big data environment | |
CN104951807A (en) | Stock market emotion determining method and device | |
CN113064994A (en) | Conference quality evaluation method, device, equipment and storage medium | |
CN114639152A (en) | Multi-modal voice interaction method, device, equipment and medium based on face recognition | |
CN104951434A (en) | Brand emotion determining method and device | |
CN109657527B (en) | Painting brush touch identification system and method | |
CN109447171A (en) | A kind of vehicle attitude classification method based on deep learning | |
CN109840537A (en) | A kind of image multitask classification method based on cross validation's neural network | |
CN106874418A (en) | A kind of automobile model data set method for building up for serving deep learning |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170620 |
|
RJ01 | Rejection of invention patent application after publication |