92 research outputs found
3D Object Retrieval using an Efficient and Compact Hybrid Shape Descriptor
No full textAbstract We present a novel 3D object retrieval method that relies upon a hybrid descriptor which is composed of 2D features based on depth buffers and 3D features based on spherical harmonics. To compensate for rotation, two alignment methods, namely CPCA and NPCA, are used while compactness is supported via scalar feature quantization to a set of values that is further compressed using Huffman coding. The superior performance of the proposed retrieval methodology is demonstrated through an extensive comparison against state-of-the-art methods on standard datasets.Eurographics 2008 Workshop on 3D Object Retrieva
*Corresponding author
No full textAbstract: This paper presents a novel unsupervised strategy for content-based image retrieval. It is based on a meaningful segmentation procedure that can provide proper distributions for matching via the Earth mover's distance as a similarity metric. The segmentation procedure is based on a hierarchical watershed-driven algorithm that extracts meaningful regions automatically. In this framework, the proposed robust feature extraction and the many-to-many region matching along with the novel region weighting for enhancing feature discrimination play a major role. Experimental results demonstrate the performance of the proposed strategy
Recognition of Urban Sound Events Using Deep Context-Aware Feature Extractors and Handcrafted Features
No full textPart 2: 8th Mining Humanistic Data WorkshopInternational audienceThis paper proposes a method for recognizing audio events in urban environments that combines handcrafted audio features with a deep learning architectural scheme (Convolutional Neural Networks, CNNs), which has been trained to distinguish between different audio context classes. The core idea is to use the CNNs as a method to extract context-aware deep audio features that can offer supplementary feature representations to any soundscape analysis classification task. Towards this end, the CNN is trained on a database of audio samples which are annotated in terms of their respective “scene” (e.g. train, street, park), and then it is combined with handcrafted audio features in an early fusion approach, in order to recognize the audio event of an unknown audio recording. Detailed experimentation proves that the proposed context-aware deep learning scheme, when combined with the typical handcrafted features, leads to a significant performance boosting in terms of classification accuracy. The main contribution of this work is the demonstration that transferring audio contextual knowledge using CNNs as feature extractors can significantly improve the performance of the audio classifier, without need for CNN training (a rather demanding process that requires huge datasets and complex data augmentation procedures)
Component Analysis
No full textcomponents, saliency Abstract: A method is proposed for constructing salient features from a set of fea-tures that are given as input to a feedforward neural network used for supervised learning. Combinations of the original features are formed that maximize the sensi-tivity of the network’s outputs with respect to variations of its inputs. The method exhibits some similarity to Principal Component Analysis, but also takes into account supervised character of the learning task. It is applied to classification problems lead-ing to improved generalization ability originating from the alleviation of the curse of dimensionality problem. This paper has not been submitted elsewhere in identical or similar form, nor will it be during the first three months after its submission to Neural Processing Letters
On the relation between discriminant analysis and mutual information for supervised linear feature extraction.
No full textAbstract This paper provides a unifying view of three discriminant linear feature extraction methods: linear discriminant analysis, heteroscedastic discriminant analysis and maximization of mutual information. We propose a model-independent reformulation of the criteria related to these three methods that stresses their similarities and elucidates their di erences. Based on assumptions for the probability distribution of the classiÿcation data, we obtain su cient conditions under which two or more of the above criteria coincide. It is shown that these conditions also su ce for Bayes optimality of the criteria. Our approach results in an information-theoretic derivation of linear discriminant analysis and heteroscedastic discriminant analysis. Finally, regarding linear discriminant analysis, we discuss its relation to multidimensional independent component analysis and derive suboptimality bounds based on information theory
Aeolus Ocean -- A simulation environment for the autonomous COLREG-compliant navigation of Unmanned Surface Vehicles using Deep Reinforcement Learning and Maritime Object Detection
No full textHeading towards navigational autonomy in unmanned surface vehicles (USVs) in
the maritime sector can fundamentally lead towards safer waters as well as
reduced operating costs, while also providing a range of exciting new
capabilities for oceanic research, exploration and monitoring. However,
achieving such a goal is challenging. USV control systems must, safely and
reliably, be able to adhere to the international regulations for preventing
collisions at sea (COLREGs) in encounters with other vessels as they navigate
to a given waypoint while being affected by realistic weather conditions,
either during the day or at night. To deal with the multitude of possible
scenarios, it is critical to have a virtual environment that is able to
replicate the realistic operating conditions USVs will encounter, before they
can be implemented in the real world. Such "digital twins" form the foundations
upon which Deep Reinforcement Learning (DRL) and Computer Vision (CV)
algorithms can be used to develop and guide USV control systems. In this paper
we describe the novel development of a COLREG-compliant DRL-based collision
avoidant navigational system with CV-based awareness in a realistic ocean
simulation environment. The performance of the trained autonomous Agents
resulting from this approach is evaluated in several successful navigations to
set waypoints in both open sea and coastal encounters with other vessels. A
binary executable version of the simulator with trained agents is available at
https://github.com/aavek/Aeolus-OceanComment: 22 pages, last blank page, 17 figures, 1 table, color, high
resolution figure
A 3D Face Recognition Algorithm Using Histogram-based Features
No full textWe present an automatic face recognition approach, which relies on the analysis of the three-dimensional facial surface. The proposed approach consists of two basic steps, namely a precise fully automatic normalization stage followed by a histogram-based feature extraction algorithm. During normalization the tip and the root of the nose are detected and the symmetry axis of the face is determined using a PCA analysis and curvature calculations. Subsequently, the face is realigned in a coordinate system derived from the nose tip and the symmetry axis, resulting in a normalized 3D model. The actual region of the face to be analyzed is determined using a simple statistical method. This area is split into disjoint horizontal subareas and the distribution of depth values in each subarea is exploited to characterize the face surface of an individual. Our analysis of the depth value distribution is based on a straightforward histogram analysis of each subarea. When comparing the feature vectors resulting from the histogram analysis we apply three different similarity metrics. The proposed algorithm has been tested with the FRGC v2 database, which consists of 4950 range images. Our results indicate that the city block metric provides the best classification results with our feature vectors. The recognition system achieved an equal error rate of 5.89% with correctly normalized face models.Eurographics 2008 Workshop on 3D Object Retrieva
- …
