Data filtering method and system for high-dimensional image data retrieval
A technology of image data and filtering method, which is applied in the field of image processing, can solve the problems of not establishing a sampling index and the disaster of dimensionality, etc., and achieve the effect of improving efficiency
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0065] like figure 1 As shown, step 100 is executed to generate a set D of high-dimensional image data points. The range of high-dimensional image data point set D is D={d i |d i ∈D}, where, d i is a high-dimensional image data point, i∈N, and N represents the number of high-dimensional image data points.
[0066] Step 110 is executed to select a reference point set f from the high-dimensional image data point set. The range of reference point set f is f={f k | f k ∈F}, where f k is the reference point, k∈K, K represents the number of the reference point. like Figure 1A As shown, in this step, step 111 is executed to randomly select a data point in the high-dimensional image data points as the first reference point f 1 . Execute step 112, select the first reference point f from the remaining high-dimensional image data points 1 The farthest point is the second reference point f 2 . Execute step 113, select f from the remaining high-dimensional image data points ...
Embodiment 2
[0071] like figure 2 As shown, a data filtering system for high-dimensional image data retrieval includes a set generation module 200 , a selection module 210 , a model generation module 220 , a filtering module 230 and an output module 240 .
[0072] Set generating module 200: for generating a set D of high-dimensional image data points. The range of high-dimensional image data point set D is D={d i |d i ∈D}, where, d i is a high-dimensional image data point, i∈N, and N represents the number of high-dimensional image data points.
[0073] Selection module 210: for selecting a reference point set F from the high-dimensional image data point set. The range of reference point set F is F={f k | f k ∈F}, where f k is the reference point, k∈K, K represents the number of the reference point. The selection method of the selection module 210 includes the following steps: Step 11: Randomly select a data point in the high-dimensional image data points as the first reference poi...
Embodiment 3
[0078] The invention proposes a data filtering method oriented to high-dimensional image data search, which performs effective pruning operation on the data to be compared, and greatly improves the search efficiency.
[0079] The implementation method is as follows:
[0080] 1. The set of data points is D, D={d i |d i ∈D}, i∈N, where N represents the number of data in the dataset.
[0081] 2. Select the reference point set F in the global data set D, F={f k | f k ∈F}, k∈K, K represents the number of reference points.
[0082] First randomly select a data point among the data points as the first reference point f 1 ;
[0083] Select and f among the remaining data points 1 The farthest point is used as the second reference point f 2 ;
[0084] Among the remaining data points select with f 1 and f 2 The distance and maximum data point is used as the new reference point;
[0085]Repeat the above process until enough datum points are selected.
[0086] 3. Calculate the...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More