Shengfeng He, Rynson W. H, Lau Wenxi Liu et al. · 2016
Abstract Existing computational models for salient object detection primarily rely on hand-crafted fea-tures, which are only able to capture low-level contrast information. In this paper, we learn the hierarchical contrast features by formulating salient object detec-tion as a binary labeling problem using deep learning techniques. A novel superpixelwise convolutional neu-ral network approach, called SuperCNN, is proposed to learn the internal representations of saliency in an effi-cient manner. In contrast to the classical convolutional networks, SuperCNN has four main properties. First, the proposed method is able to learn the hierarchical contrast features, as it is fed by two meaningful super-pixel sequences, which is much more effective for detect-ing salient regions than feeding raw image pixels. Sec-ond, as SuperCNN recovers the contextual information among superpixels, it enables large context to be in-volved in the analysis efficiently. Third, benefiting from the superpixelwise mechanism, the required number of predictions for a densely labeled map is hugely reduced. Fourth, saliency can be detected independent of region size by utilizing a multiscale network structure. Experi-ments show that SuperCNN can robustly detect salient objects and outperforms the state-of-the-art methods on three benchmark datasets.
Cited by 0
A. C. Faul · 2025
Cited by 0
Jennifer Louie, Jennifer Louie, Jennifer Louie · 2003
Previous biological models of object recognition in cortex have been evaluated using idealized scenes and have hard-coded features, such as the HMAX model by Riesenhuber and Poggio [10]. Because HMAX uses the same set of features for all object classes, it does not perform well in the task of detecting a target object in clutter. This thesis presents a new model that integrates learning of object-specific features with the HMAX. The new model performs better than the standard HMAX and comparably to a computer vision system on face detection. Results from experimenting with unsupervised learning of features and the use of a biologically-plausible classifier are presented.
Cited by 0
Laux Darmstadt, Melvin Laux Darmstadt, Tag Der Einreichung · 2014
Hiermit versichere ich, die vorliegende Bachelor-Thesis ohne Hilfe Dritter nur mit den angegebenen Quellen und Hilfsmitteln angefertigt zu haben. Alle Stellen, die aus Quellen entnommen wurden, sind als solche kenntlich gemacht. Diese Arbeit hat in gleicher oder ähnlicher Form noch keiner Prüfungs-behörde vorgelegen. Darmstadt, den 10. Oktober 2014 (Melvin Laux) High Dimensional Reinforcement Learning requires a useful low dimensional representation of the input data to be able to train an agent. However, learning these features requires complete knowledge about the input data, which in general is not available. Initially, to acquire the needed knowledge, many experiments have to be used to explore the input data space at the before training. Preferably, the process of learning useful features should be done on the fly, during the training of the agent. One major challenge for this online feature learning is to find a feature learning method, that is time-efficient and able to adapt to changing input data distributions. One method that meets both of these criteria are incremental autoencoders. In this thesis, we evaluate the perormance of incremental autoencoders for online feature learning for Q-Learning and present limited results of applying incremental autoencoders as feature learning method on the visual pole balancing task. Zusammenfassung
Cited by 0
Behzad Ghazanfari, Fatemeh Afghah, MohammadTaghi Hajiaghayi · IEEE Access · 2020
This paper proposes <i>inverse feature learning</i> (IFL) as a novel supervised feature learning technique that learns a set of high-level features for classification based on an <i>error representation</i> approach. The key contribution of this method is to learn the representation of error as high-level features, while current representation learning methods interpret error by loss functions which are obtained as a function of differences between the true labels and the predicted ones. One advantage of this error representation is that the learned features for each class can be obtained independently of learned features for other classes; therefore, IFL can learn simultaneously meaning that it can learn new classes' features without retraining. Error representation learning can also help with generalization and reduce the chance of over-fitting by adding a set of impactful features to the original data set which capture the relationships between each instance and different classes through an error generation and analysis process. This method can be particularly effective in data sets, where the instances of each class have diverse feature representations or the ones with imbalanced classes. The experimental results show that the proposed IFL results in better performance compared to the state-of-the-art classification techniques for several popular data sets. We hope this paper can open a new path to utilize the proposed perspective of error representation learning in different feature learning domains.
Cited by 0Open access
Devi, T. Rupa · Figshare · 2025
The CGAN method enhances data diversity by generating synthetic training data, which helps the model generalise better and avoid overfitting. This step aids in improving the system's ability to summarise text from varied domains and ensures more robust training with limited available data.
Cited by 0Open access
Marcus Hütter, David Quarel, Elliot Catt · 2024
Cited by 0
Donahue · TIB Data Manager · 2024
Cited by 0Open access
SpringerReference · 2012
Cited by 0
Simon Hickinbotham, Edwin R. Hancock, J. Austin · 2002
This paper describes a statistical framework for the unsupervised learning of linear filter combinations for feature characterisation. The learning strategy is two step. In the first instance, the EM algorithm is used to learn the foreground probability distribution. This is an abductive process, since we have a detailed model of the background process based on the known noise-response characteristics of the filter-bank. The second phase uses the a posteriori foreground and background probabilities to compute a weighted between-class covariance matrix. We use the principal components analysis to locate the linear filter combinations that maximise the between class covariance matrix. The new feature characterisation method is illustrated for the problem of extracting linear features from complex millimetre radar images. The method proves to be effective in learning a mixture of sine and cosine phase Gabor functions necessary to capture shadowed line structures.
Cited by 1
Gabriela Rosenblau, Jonathan K Doyon, Sarah Shomstein et al. · OSF Preprints (OSF Preprints) · 2025
Cited by 0
Björn Lindström · OSF Preprints (OSF Preprints) · 2024
Cited by 0
Y-Lan Boureau · 2012
Cited by 0
Devi, T. Rupa · Figshare · 2025
The CGAN method enhances data diversity by generating synthetic training data, which helps the model generalise better and avoid overfitting. This step aids in improving the system's ability to summarise text from varied domains and ensures more robust training with limited available data.
Cited by 0Open access
Behzad Ghazanfari, Fatemeh Afghah, MohammadTaghi Hajiaghayi · arXiv (Cornell University) · 2020
This paper proposes inverse feature learning as a novel supervised feature\nlearning technique that learns a set of high-level features for classification\nbased on an error representation approach. The key contribution of this method\nis to learn the representation of error as high-level features, while current\nrepresentation learning methods interpret error by loss functions which are\nobtained as a function of differences between the true labels and the predicted\nones. One advantage of such learning method is that the learned features for\neach class are independent of learned features for other classes; therefore,\nthis method can learn simultaneously meaning that it can learn new classes\nwithout retraining. Error representation learning can also help with\ngeneralization and reduce the chance of over-fitting by adding a set of\nimpactful features to the original data set which capture the relationships\nbetween each instance and different classes through an error generation and\nanalysis process. This method can be particularly effective in data sets, where\nthe instances of each class have diverse feature representations or the ones\nwith imbalanced classes. The experimental results show that the proposed method\nresults in significantly better performance compared to the state-of-the-art\nclassification techniques for several popular data sets. We hope this paper can\nopen a new path to utilize the proposed perspective of error representation\nlearning in different feature learning domains.\n
Cited by 0Open access
Tobias Schlagenhauf, Yiwen Lin, Benjamin Noack · Machine Vision and Applications · 2023
Abstract Ensembles of convolutional neural networks have shown remarkable results in learning discriminative semantic features for image classification tasks. However, the models in the ensemble often concentrate on similar regions in images. This work proposes a novel method that forces a set of base models to learn different features for a classification task. These models are combined in an ensemble to make a collective classification. The key finding is that by forcing the models to concentrate on different features, the classification accuracy is increased. To learn different feature concepts, a so-called feature distance loss is implemented on the feature maps. The experiments on benchmark convolutional neural networks (VGG16, ResNet, AlexNet), popular datasets (Cifar10, Cifar100, miniImageNet, NEU, BSD, TEX), and different training samples (3, 5, 10, 20, 50, 100 per class) show the effectiveness of the proposed feature loss. The proposed method outperforms classical ensemble versions of the base models. The Class Activation Maps explicitly prove the ability to learn different feature concepts. The code is available at: https://github.com/2Obe/Feature-Distance-Loss.git .
Cited by 5Open access
Marwa Farouk Ibrahim Ibrahim, Adel Al-Jumaily · 2016
Feature extraction is playing a major role in bio signal processing. Feature identification and selection has two approaches. The common approach is engineering handcraft which is based on user experience and application area. While the other approach is feature learning that based on making the system identify and select the best features suit the application. The idea behind feature learning is to avoid dealing with any feature extraction or reduction algorithms and to train the suggested model on learning with avoiding the exposure to feature extraction which is mainly based on researcher experience. In this paper, Independent component analysis (ICA) will be implemented as a feature learning technique to learn the model extract the features from the input data. Deep learning approach will be proposed by implementing ICA to learn features. In the proposed model, the raw data will be read then represented by using different signal representation as Spectrogram, Wavelet and Wavelet Packet. Then, the new represented data will be fed to Independent component analysis layer to generate features and finally, the performance of the suggested scheme will be evaluated by applying different classifiers such as Support Vector Machine, Extreme Learning Machine and Discriminate Analysis. And As an improving step for the results, classifier fusion layer will be implemented to select the most accurate result for both training and testing set. Classifier fusion layer resulted in a promising training and testing accuracies. On the other side, Feature Selection is the process of selecting subset of features from a pool of features.
Cited by 3
Laurent Thomas, James H. von Brecht, Xavier Bresson · arXiv (Cornell University) · 2022
We propose a simple data model inspired from natural data such as text or images, and use it to study the importance of learning features in order to achieve good generalization. Our data model follows a long-tailed distribution in the sense that some rare subcategories have few representatives in the training set. In this context we provide evidence that a learner succeeds if and only if it identifies the correct features, and moreover derive non-asymptotic generalization error bounds that precisely quantify the penalty that one must pay for not learning features.
Cited by 1Open access
Shuai Yang, Xin Li, Ming-Hung Wu et al. · IEEE Transactions on Big Data · 2024
Stable learning aims to leverage the knowledge in a relevant source domain to learn a prediction model that can generalize well to target domains. Recent advances in stable learning mainly proceed by eliminating spurious correlations between irrelevant features and labels through sample reweighting or causal feature selection. However, most existing stable learning methods either only weaken partial spurious correlations or discard part of true causal relationships, resulting in generalization performance degradation. To tackle these issues, we propose the Dual Feature Learning (DFL) algorithm for stable learning, which consists of two phases. Phase 1 first learns a set of sample weights to balance the distribution of treated and control groups corresponding to each feature, and then uses the learned sample weights to assist feature selection to identify part of irrelevant features for completely isolating spurious correlations between these irrelevant features and labels. Phase 2 first learns two groups of sample weights again using the subdataset after feature selection, and then obtains high-quality feature representations by integrating a weighted cross-entropy model and an autoencoder model to further get rid of spurious correlations. Using synthetic and four real-world datasets, the experiments have verified the effectiveness of DFL, in comparison with eleven state-of-the-art methods.
Cited by 0
Chamara Kasun Liyanaarachchi Lekamalage · 2017
ola Jones algorithm locates the face in the image; 2) LRF-ELMAE estimates the facial points on the located faces using mean face shape; 3) Extreme Learning Machine Supervised Descent Method (ELM-SDM) improves the estimated facial points using low dimensional response maps. The experimental results show that ELM-FPD outperforms other multi-layer neural networks for facial point detection.
Cited by 0Open access