Axel Pinz

Multispectral classification of Landsat-images using neural networks

The authors report the application of three-layer back-propagation networks for classification of Landsat TM data on a pixel-by-pixel basis. The results are compared to Gaussian maximum likelihood classification. First, it is shown that the neural network is able to perform better than the maximum likelihood classifier. Secondly, in an extension of the basic network architecture it is shown that textural information can be integrated into the neural network classifier without the explicit definition of a texture measure. Finally, the use of neural networks for postclassification smoothing is examined. >

Generic object recognition with boosting

This paper explores the power and the limitations of weakly supervised categorization. We present a complete framework that starts with the extraction of various local regions of either discontinuity or homogeneity. A variety of local descriptors can be applied to form a set of feature vectors for each local region. Boosting is used to learn a subset of such feature vectors (weak hypotheses) and to combine them into one final hypothesis for each visual category. This combination of individual extractors and descriptors leads to recognition rates that are superior to other approaches which use only one specific extractor/descriptor setting. To explore the limitation of our system, we had to set up new, highly complex image databases that show the objects of interest at varying scales and poses, in cluttered background, and under considerable occlusion. We obtain classification results up to 81 percent ROC-equal error rate on the most complex of our databases. Our approach outperforms all comparable solutions on common databases.

A boundary-fragment-model for object detection

The objective of this work is the detection of object classes, such as airplanes or horses. Instead of using a model based on salient image fragments, we show that object class detection is also possible using only the objects boundary. To this end, we develop a novel learning technique to extract class-discriminative boundary fragments. In addition to their shape, these “codebook” entries also determine the objects centroid (in the manner of Leibe et al. [19]). Boosting is used to select discriminative combinations of boundary fragments (weak detectors) to form a strong “Boundary-Fragment-Model” (BFM) detector. The generative aspect of the model is used to determine an approximate segmentation. We demonstrate the following results: (i) the BFM detector is able to represent and detect object classes principally defined by their shape, rather than their appearance; and (ii) in comparison with other published results on several object classes (airplanes, cars-rear, cows) the BFM detector is able to exceed previous performances, and to achieve this with less supervision (such as the number of training images).

Weak Hypotheses and Boosting for Generic Object Detection and Recognition

In this paper we describe the first stage of a new learning system for object detection and recognition. For our system we propose Boosting (5) as the underlying learning technique. This allows the use of very diverse sets of visual features in the learning process within a com- mon framework: Boosting — together with a weak hypotheses finder — may choose very inhomogeneous features as most relevant for combina- tion into a final hypothesis. As another advantage the weak hypotheses finder may search the weak hypotheses space without explicit calculation of all available hypotheses, reducing computation time. This contrasts the related work of Agarwal and Roth (1) where Winnow was used as learning algorithm and all weak hypotheses were calculated explicitly. In our first empirical evaluation we use four types of local descriptors: two basic ones consisting of a set of grayvalues and intensity moments and two high level descriptors: moment invariants (8) and SIFTs (12). The descriptors are calculated from local patches detected by an inter- est point operator. The weak hypotheses finder selects one of the local patches and one type of local descriptor and efficiently searches for the most discriminative similarity threshold. This differs from other work on Boosting for object recognition where simple rectangular hypotheses (22) or complex classifiers (20) have been used. In relatively simple images, where the objects are prominent, our approach yields results comparable to the state-of-the-art (3). But we also obtain very good results on more complex images, where the objects are located in arbitrary positions, poses, and scales in the images. These results indicate that our flexible approach, which also allows the inclusion of features from segmented re- gions and even spatial relationships, leads us a significant step towards generic object recognition.

Robust Pose Estimation from a Planar Target

In theory, the pose of a calibrated camera can be uniquely determined from a minimum of four coplanar but noncollinear points. In practice, there are many applications of camera pose tracking from planar targets and there is also a number of recent pose estimation algorithms which perform this task in real-time, but all of these algorithms suffer from pose ambiguities. This paper investigates the pose ambiguity for planar targets viewed by a perspective camera. We show that pose ambiguities - two distinct local minima of the according error function - exist even for cases with wide angle lenses and close range targets. We give a comprehensive interpretation of the two minima and derive an analytical solution that locates the second minimum. Based on this solution, we develop a new algorithm for unique and robust pose estimation from a planar target. In the experimental evaluation, this algorithm outperforms four state-of-the-art pose estimation algorithms

Incremental learning of object detectors using a visual shape alphabet

We address the problem of multiclass object detection. Our aims are to enable models for new categories to benefit from the detectors built previously for other categories, and for the complexity of the multiclass system to grow sublinearly with the number of categories. To this end we introduce a visual alphabet representation which can be learnt incrementally, and explicitly shares boundary fragments (contours) and spatial configurations (relation to centroid) across object categories. We develop a learning algorithm with the following novel contributions: (i) AdaBoost is adapted to learn jointly, based on shape features; (ii) a new learning schedule enables incremental additions of new categories; and (iii) the algorithm learns to detect objects (instead of categorizing images). Furthermore, we show that category similarities can be predicted from the alphabet. We obtain excellent experimental results on a variety of complex categories over several visual aspects. We show that the sharing of shape features not only reduces the number of features required per category, but also often improves recognition performance, as compared to individual detectors which are trained on a per-class basis.

Comparison of 4 methods for quantifying posterior capsule opacification.

Purpose: To compare the results of posterior capsule opacification (PCO) quantification and the repeatability of a fully automated analysis system (Automated Quantification of After‐Cataract [AQUA]) with that of 2 other quantification methods and subjective grading of PCO. A test set of digital retroillumination images of 100 eyes with PCO of varying degrees was used. Setting: Department of Ophthalmology, University of Vienna, Vienna, Austria. Methods: One hundred digital retroillumination images of eyes (100 patients) with PCO were selected to attain an even distribution from mild to severe cases. The images were evaluated by 4 methods: subjective grading by 4 experienced and 4 inexperienced examiners, the subjective Evaluation of Posterior Capsular Opacification (EPCO) system, posterior capsule opacification (POCO) software, and the AQUA system. Ten images were presented twice to assess the reproducibility of the analysis systems. Results: Subjective grading correlated best with the subjective EPCO system and the objective AQUA system (r = 0.94 and r = 0.93, respectively). The POCO system showed very early saturation and therefore a much weaker correlation (r = 0.73). The POCO scores reached the maximum of 100% in several minimal to mild PCO cases. The reproducibility of the AQUA software was perfect and that of the other analysis systems, comparably satisfactory. Conclusion: The objective AQUA score correlated well with subjective methods including the EPCO system. The POCO system, which assesses PCO area, did not adequately describe PCO intensity and includes a subjective step in the analysis process. The AQUA system could become an important tool for randomized masked trials of PCO inhibition.

Appearance-based active object recognition

We present an efficient method within an active vision framework for recognizing objects which are ambiguous from certain viewpoints. The system is allowed to reposition the camera to capture additional views and, therefore, to improve the classification result obtained from a single view. The approach uses an appearance-based object representation, namely the parametric eigenspace, and augments it by probability distributions. This enables us to cope with possible variations in the input images due to errors in the pre-processing chain or changing imaging conditions. Furthermore, the use of probability distributions gives us a gauge to perform view planning. Multiple observations lead to a significant increase in recognition rate. Action planning is shown to be of great use in reducing the number of images necessary to achieve a certain recognition performance when compared to a random strategy. q 2000 Elsevier Science B.V. All rights reserved.

Active object recognition by view integration and reinforcement learning

Abstract A mobile agent with the task to classify its sensor pattern has to cope with ambiguous information. Active recognition of three-dimensional objects involves the observer in a search for discriminative evidence, e.g., by change of its viewpoint. This paper defines the recognition process as a sequential decision problem with the objective to disambiguate initial object hypotheses. Reinforcement learning provides then an efficient method to autonomously develop near-optimal decision strategies in terms of sensorimotor mappings. The proposed system learns object models from visual appearance and uses a radial basis function (RBF) network for a probabilistic interpretation of the two-dimensional views. The information gain in fusing successive object hypotheses provides a utility measure to reinforce actions leading to discriminative viewpoints. The system is verified in experiments with 16 objects and two degrees of freedom in sensor motion. Crucial improvements in performance are gained using the learned in contrast to random camera placements.

Learning an Alphabet of Shape and Appearance for Multi-Class Object Detection

We present a novel algorithmic approach to object categorization and detection that can learn category specific detectors, using Boosting, from a visual alphabet of shape and appearance. The alphabet itself is learnt incrementally during this process. The resulting representation consists of a set of category-specific descriptors—basic shape features are represented by boundary-fragments, and appearance is represented by patches—where each descriptor in combination with centroid vectors for possible object centroids (geometry) forms an alphabet entry. Our experimental results highlight several qualities of this novel representation. First, we demonstrate the power of purely shape-based representation with excellent categorization and detection results using a Boundary-Fragment-Model (BFM), and investigate the capabilities of such a model to handle changes in scale and viewpoint, as well as intra- and inter-class variability. Second, we show that incremental learning of a BFM for many categories leads to a sub-linear growth of visual alphabet entries by sharing of shape features, while this generalization over categories at the same time often improves categorization performance (over independently learning the categories). Finally, the combination of basic shape and appearance (boundary-fragments and patches) features can further improve results. Certain feature types are preferred by certain categories, and for some categories we achieve the lowest error rates that have been reported so far.