Separation method for multi-object detection in digital images
A digital image and separation method technology, applied in the field of image processing, can solve problems such as deteriorating detection results and slow calculation speed, and achieve the effect of reducing complexity and improving accuracy
Inactive Publication Date: 2013-03-27
SHANGHAI JIAO TONG UNIV
View PDF2 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Therefore, this algorithm may produce degraded detection results
At the same time, during the training process of the spatial model, the structured support vector machine needs to optimize all objects in all pictures to get the final model, so the calculation speed is slow
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
example
[0052] The image database used in this embodiment is all images of PASCAL VOC 2007. PASCAL VOC is an authoritative competition in the fields of image classification, image detection, and gesture recognition. All object categories in the database are used, namely: aeroplane, bicycle, bird, boat, bottle, bus, car, cat, chair, cow, dining table, dog, horse, motorbike, person, potted plant, sheep, sofa , train, TV monitor, a total of 20 categories. The entire image database has a total of 10,000 images, of which the training set and the test set each account for half. On this dataset, the present embodiment greatly outperforms other object detection methods. For comparison, the detection model (Def) using local features, the detection model (SSVM) using structured support vector machine, the detection model (Thr) using threshold method, and the detection model using maximum value method and expanding spatial features are given here. Detection model (Max). The above four method...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a separation method for multi-object detection in digital images, and belongs to the technical field of image processing. The separation method includes steps of (1), detecting a digital image by the aid of local features, and separating detection windows from one another; and (2), creating and extending spatial features according to separated window information, selecting optimal parameters by means of cross validation, and performing training and testing by the aid of a support vector machine. The separation method has the advantages that spatial relations among objects in the image are effectively utilized, and the computational complexity is reduced; as the windows are separated from one another, all detection results cannot be deteriorated by few detection errors; large quantities of features of the spatial relations can be introduced, and accordingly the object detection results are effectively improved; and the separation method is superior to various existing object detection methods.
Description
technical field [0001] The invention relates to a multi-object detection method, in particular to a separation method for multi-object detection in a digital image, and belongs to the technical field of image processing. Background technique [0002] In recent years, with the rapid development of multimedia technology and computer network, massive digital images are generated and disseminated every day. Automatically identifying the content of these digital images, so as to efficiently organize and manage them, has become a hot and difficult point of current research. Object detection in digital images is one of the key technologies to solve image content recognition. In traditional object detection schemes, many are based on local features of the image, such as image color, shape, edge texture, etc., without considering the relationship between objects in the image. This method relies on the invariance of local features, and is less effective in detecting objects with var...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/46
Inventor 张瑞朱玉琨朱俊邹维嘉仇媛媛付赛男
Owner SHANGHAI JIAO TONG UNIV