IEEE Transactions on Pattern Analysis and Machine Intelligence
Digital Image Processing Using MATLAB
Digital Image Processing Using MATLAB
Distinctive Image Features from Scale-Invariant Keypoints
International Journal of Computer Vision
A tutorial on support vector regression
Statistics and Computing
Neighborhood Preserving Embedding
ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories
CVPR '06 Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Volume 2
Automatic recognition of biological particles in microscopic images
Pattern Recognition Letters
AAAI'07 Proceedings of the 22nd national conference on Artificial intelligence - Volume 1
PCA-SIFT: a more distinctive representation for local image descriptors
CVPR'04 Proceedings of the 2004 IEEE computer society conference on Computer vision and pattern recognition
Learning Linear Discriminant Projections for Dimensionality Reduction of Image Descriptors
IEEE Transactions on Pattern Analysis and Machine Intelligence
SURF: speeded up robust features
ECCV'06 Proceedings of the 9th European conference on Computer Vision - Volume Part I
Hi-index | 0.00 |
The microscopic analysis of the urinary sediment is important in making diagnoses for a variety of diseases, including urinary tract infection, urinary tract tumors, occult glomerulonephritis, and interstitial nephritis. A typical automated system acquires images of urinary sediment by employing a CCD camera, and then detects and recognizes the distinct particles automatically from these images. Automated recognition of these particles represents a significant challenge due to poor image resolution, strong variability of particle shape and size, and challenges associated with detection of particles in the presence of noisy backgrounds. In this paper, we present a novel method for urine particle classification based on the use of local descriptors coupled with regression based decision fusion. Specifically, DAISY descriptors have been used to capture the textural characteristics of each particle and subjected to dimensionality reduction across three linear subspaces to increase the diversity in decision making along with lowering the "curse of dimensionality". Classification in each subspace is based on computing a similarity score, which is then fused through support vector regression to obtain a final classification. The approach is applied to both brightfield and multispectral data to ascertain the benefits of multispectral imaging for urine analysis. Urine particles analyzed included crystals, casts and blood cells, and the results obtained show an average classification accuracy of 92.6% for 6 classes of urinary particles.