Inna Stainvas

Learn More
It is well known that even slight changes in nonuniform illumination lead to a large image variability and are crucial for many visual tasks. This paper presents a new ICA related probabilistic model where the number of sources exceeds the number of sensors to perform an image segmentation and illumination removal, simultaneously. We model illumination and(More)
We propose a novel multiple-instance learning (MIL) algorithm for designing classifiers for use in computer aided detection (CAD). The proposed algorithm has 3 advantages over classical methods. First, unlike traditional learning algorithms that minimize the candidate level misclassification error, the proposed algorithm directly optimizes the patient-wise(More)
This Letter addresses image segmentation via a generative model approach. A Bayesian network (BNT) in the space of dyadic wavelet transform coefficients is introduced to model texture images. The model is similar to a Hidden Markov model (HMM), but with non-stationary transitive conditional probability distributions. It is composed of discrete hidden(More)
Learning a many-parameter model is generally an under-constrained problem that requires additional regularization. We study several information theoretic constraints and show that reconstruction constraints achieve improved performance. In addition to studying entropy and BCM constraints on the hidden units of a feed-forward architecture, we derive a new(More)
Learning a many-parameter model is generally an under-constrained problem that requires additional regularization. We propose to use reconstruction as a regularization constraint for image classification. We show that fusing the two models together is an effective regularizer which adds to the improvement achieved by weight decay constraints. This(More)
Curbs are important cues identifying the boundary of a roadway. Their detection is required by many automotive features. Currently there is no an agreed benchmark to report and compare curb detection results. This paper presents annotation and performance evaluation toolboxes (PET) developed by us for measuring the performance of curb detection algorithms.(More)