Loading...
Please wait, while we are loading the content...
Similar Documents
A logistic regression approach to content-based mammogram retrieval
| Content Provider | Semantic Scholar |
|---|---|
| Author | Wei, Chia-Hung |
| Copyright Year | 2006 |
| Abstract | Content-based image retrieval (CBIR) has been proposed to address the problem of image retrieval from medical image databases. Relevance feedback, explaining the user's query concept, can be used to bridge the semantic gap and improve the performance of CBIR systems. This paper proposes a learning method for relevance feedback, which develops logistic regression models to generalize the 2-class problem and provide an estimate of probability of class membership. To build the model, relevance feedback is used as the training data and the iteratively re-weighted least squares method is applied to estimate the parameters of the regression curve and compute the maximum likelihood. After logistic regression models are fitted, discriminating features are selected by the measure of goodness of fit statistics. The weights of those discriminating features are determined based on their individual contributions to the maximum likelihood. The probability of class membership can therefore be obtained for each image of the database. Experimental results show that the proposed learning method can effectively improve the average precision from 41% to 63% through five iterations of relevance feedback rounds. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://wrap.warwick.ac.uk/61497/1/WRAP_Li_cs-rr-426.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
