Water-Level model is an effective method in density-based classification. We use biased sampling, local similarity and popularity as preprocessing, and employ a merging operation in the water-level model for classification. Biased sampling is to get some information about the global structure. Similarity and local density are mainly used to understand the local structure. In biased sampling, images are divided into many l x l patches and a sample pixel is selected from each patch. Similarity at a point p, denoted by sim(p), measures the change of gray level between point p and its neighborhood N(p). Besides using biased sampling to combine spectral and spatial information, we use similarity and local popularity in selecting sample points. A sample point is chosen based on the minimum value of sim(p) + [1-P(p)] after normalization. The selected pixel is a better representative, especially near the border of an object. To make it more effective, one has to deal with small spikes and bumps. To get rid of the small spikes, we establish a threshold |[f(P1)-f(P2)]*(P1-P2)| > c*l*l , where c is a constant, P1 is a local maximum point to be tested and P2 is the nearest local minimum from P1. The condition is only related to the size of the patches l*l. The merging operation we include in the model makes the threshold constant less sensitive in the process. DBScan is combined with the enhanced water level model to reduce noise and to get connected components. Preliminary experiments have been conducted using the proposed methods and the results are promising.
Water level model is an effective method in density-based classification. To improve the result, we use biased sampling, local similarity and popularity as preprocessing, and then apply the water-level model for classification. Biased sampling is to get some information about the global structure. Similarity and local density are mainly used to understand the local structure. In biased sampling, images are divided into many l*l patches and a sample pixel is selected from each patch. Similarity at a point p, denoted by sim(p), measures the change of gray level between point p an its neighborhood N(p). Besides using biased sampling to combine spectral and spatial information, we use similarity and local popularity in selecting sample points. A sample point is chosen based on the minimum value of sin(p) + [1-P(p)] after normalization. The selected pixel is a better representative, especially near the border of an object. Kernel estimators are employed to obtain smooth density approximation. The water-level model is relatively easy and effective when the density function is smoothed. To make it more effective in other cases, one has to deal with small spikes and bumps. To get rid of the small spikes, we establish a threshold ê[f(P1) - f(P 2)*(P1-P 2) ê > c*l*l , where c is a constant, P1 is a local maximum point to be tested and P2 is the nearest local minimum form P1. The condition is only related to the size of the patches l*l. After using the average filter, we choose l to be the square root of the fifth peak if it is between 5 and 20, otherwise set l = 10. Preliminary experiments have been conducted using proposed methods with different values of the constant c in the threshold condition. Experimental results are provided.
Human interpreters are very sensitive to spatial information in supervised classification. A well-known isodata algorithm in unsupervised classification requires many parameters to be set by human being. Some other unsupervised algorithm focuses on spectral information, but spatial information is lost in the process. Biased sampling is one good approach to get some information about the global structure. For local structures, many techniques have been used. For example similarity and local density are discussed in many papers. In biased sampling, images are divided into many lxl patches and a sample pixel is selected from each patch. Similarity at a point p, denoted by sim(p), measures the change of gray level between point p and its neighborhood N(p). In this article we introduce a method to use biased sampling to combine spectral and spatial information. We use similarity and local popularity in selecting sample points to get better results. To use similarity (sim(p)≤δ), one must determine δ. One way is to make it adapted such that a sample point can be selected from each patch. Here after normalization, we choose a sample point with a minimum value of [equation] for some positive numbers α and β. There is no precondition for δ needed and the selected pixel is a better representative, especially near the border of an object. Kernel estimators are employed to obtain smooth density approximation before final classification. Some experiments have been conducted using the proposed methods and the results are satisfactory.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.