In this paper, we address fine-grained classification which is quite challenging due to high intra-class variations and subtle inter-class variations. Most modern approaches to fine-grained recognition are established based on convolutional neural networks (CNN). Despite the effectiveness, these approaches still suffer from two major problems. First, they highly rely on large sets of training data, but manually annotating numerous training data is expensive. Second, the learned feature presentations by these approaches are often of high dimensions, leading to less efficiency. To tackle the two problems, we present an approach where on-line dictionary learning is integrated into CNN. The dictionaries can be incrementally learned by leveraging a vast amount of weakly labeled data on the Internet. With these dictionaries, all the training and testing data can be sparsely represented. Our approach is evaluated and compared with the state-of-the-art approaches on the benchmark dataset, CUB-200-2011. The promising results demonstrate its superiority in both efficiency and accuracy.

Fine-grained image classification is quite challenging due to high inter-class similarity and large intra-class variations. Another issue is the small amount of training images with a large number of classes to be identified. To address the challenges, we propose a model for fine-grained image classification with its application to bird species recognition. Based on the features extracted by bilinear convolutional neural network (BCNN), we propose an on-line dictionary learning algorithm where the principle of sparsity is integrated into classification. The features extracted by BCNN encode pairwise neuron interaction in a translation-invariant manner. This property is valuable to fine-grained classification. The proposed algorithm for dictionary learning further carries out sparsity based classification, where training data can be represented with a less number of dictionary atoms. It alleviates the problems caused by insufficient training data, and makes classification much more efficient. Our approach is evaluated and compared with the state-of-the-art approaches on the CUB-200-2011 dataset. The promising experimental results demonstrate its efficacy and superiority.

Feature extraction is one of the most important phases of medical image classification which requires extensive domain knowledge. Convolutional Neural Networks (CNN) have been successfully used for feature extraction in images from different domains involving a lot of classes. In this paper, CNNs are exploited to extract a hierarchical and discriminative representation of X-ray images. This representation is then used for classification of the X-ray images as various parts of the body. Visualization of the feature maps in the hidden layers show that features learnt by the CNN resemble the essential features which help discern the discrimination among different body parts. A comparison on the standard IRMA X-ray image dataset demonstrates that the CNNs easily outperform classifiers with hand-engineered features.

Action recognition in unconstrained videos is one of the most important challenges in computer vision. In this paper, we propose sparsity-inducing dictionaries as an effective representation for action classification in videos. We demonstrate that features obtained from sparsity based representation provide discriminative information useful for classification of action videos into various action classes. We show that the constructed dictionaries are distinct for a large number of action classes resulting in a significant improvement in classification accuracy on the HMDB51 dataset. We further demonstrate the efficacy of dictionaries and sparsity based classification on other large action video datasets like UCF50.

In this paper, an approach for classification of medical images using edge-based features is proposed. We demonstrate that the edge information extracted from an image by dividing the image into patches and each patch into concentric circular regions provide discriminative information useful for classification of medical images by considering 18 categories of radiological medical images namely, skull, hand, breast, cranium, hip, cervical spin, pelvis, radiocarpaljoint, elbow etc.,. The ability of On-line Dictionary Learning (ODL) to achieve sparse representation of an image is exploited to develop dictionaries for each class using edge-based feature. A low rate of misclassification error for these test images validates the effectiveness of edge-based features and On-line Dictionary Learning models for classification of medical images.

In this paper, a clustering method using dictionary learning is proposed to group large medical databases. An approach grouping similar images into clusters that are sparsely represented by the dictionaries and learning dictionaries simultaneously via K-SVD is proposed. A query image is matched with the existing dictionaries to identify the dictionary with the sparsest representation using an Orthogonal Matching Pursuit (OMP) algorithm. Then images in the cluster associated with this dictionary are compared using a similarity measure to retrieve images similar to the query image. The main features of the method are that it requires no training data and works well on the medical databases which are not restricted to specific context. The performance of the proposed method is examined on IRMA test image database. The experimental results demonstrate the efficacy of the proposed method in the retrieval of medical images.

We propose an approach for sparse representation of dense features for action classification. Sparse representation has already been shown in literature as a good approximation for signals for various computer vision applications. This property is leveraged to represent a dense feature like action bank in the form of sparse dictionaries. These dictionaries are learnt using on-line dictionary learning (ODL) which further facilitates incorporating new training examples into existing dictionaries for more robust representation of various categories of action as and when required. Evaluation of the proposed method on realistic action datasets like UCF50 and HMDB51 shows that considering sparse representation of a dense feature is more suitable for classification than the feature itself.

Speckle is a multiplicative noise which is inherent in medical ultrasound images. Speckles contributes high variance between neighboring pixels reducing the visual quality of an image. Suppression of speckle noise significantly improves the diagnostic content present in the image. In this paper, we propose how sparseland model can be used for speckle suppression. The performance of the model is evaluated based on variance to mean ratio of a patch in the filtered image. The algorithm is tested on both software generated images and real time ultrasound images. The proposed algorithm has performed similar to past adaptive speckle suppression filters and seems promising in improving diagnostic content.

Classification of medical data is one of the most challenging pattern recognition problems. As stated in literature a single classifier is unable to solve all medical image classification problems due to high sensitivity to noise and other imperfections like data imbalance. So, several individual classifiers have been studied to solve the different types of classification problems arising in medical datasets but all have proven to be useful on some specific datasets. Hence, in this paper, we propose a generic multi-level classification approach for medical datasets using sparsity based dictionary learning and support vector machine approaches. The proposed technique demonstrates the following advantages: 1) gives better performance of classification accuracy over all datasets 2) solves imbalanced data problems 3) needs no fusion and ensemble methods in multi-level classification. The results presented on the 5 standard UCI medical datasets demonstrate that the efficacy of the proposed multi-level classification technique.

Cardiovascular diseases (CVD) are a leading cause of unnecessary hospital admissions as well as fatalities placing an immense burden on the healthcare industry. A process to provide timely intervention can reduce the morbidity rate as well as control rising costs. Patients with cardiovascular diseases require quick intervention. Towards that end, automated detection of abnormal heartbeats captured by electronic cardiogram (ECG) signals is vital. While cardiologists can identify different heartbeat morphologies quite accurately among different patients, the manual evaluation is tedious and time consuming. In this chapter, we propose new features from the time and frequency domains and furthermore, feature normalization techniques to reduce inter-patient and intra-patient variations in heartbeat cycles. Our results using the adaptive learning based classifier emulate those reported in existing literature and in most cases deliver improved performance, while eliminating the need for labeling of signals by domain experts.

The recent developments in the field of compressed sensing (CS) have been shown to have tremendous potential for applications such as content-based image retrieval. The underdetermined framework present in CS requires some implicit assumptions on the image database or needs the projection (or downsampling) of database members into lower dimensional space. The present work, however, poses the problem of image retrieval in overdetermined setting. The main feature of the proposed method is that it does not require any downsampling operation or implicit assumption on the databases. Our experimental results demonstrate that our method has potential for such applications as content-based image retrieval.

In this paper, we proposed a method for classification of medical images captured by different sensors (modalities) based on multi-scale wavelet representation using dictionary learning. Wavelet features extracted from an image provide discrimination useful for classification of medical images, namely, diffusion tensor imaging (DTI), magnetic resonance imaging (MRI), magnetic resonance angiography (MRA) and functional magnetic resonance imaging (FRMI). The ability of On-line dictionary learning (ODL) to achieve sparse representation of an image is exploited to develop dictionaries for each class using multi-scale representation (wavelets) feature. An experimental analysis performed on a set of images from the ICBM medical database demonstrates efficacy of the proposed method.

Classifying action videos became challenging problem in computer vision community. In this work, action videos are represented by dictionaries which are learned by online dictionary learning (ODL). Here, we have used two simple measures to classify action videos, reconstruction error and projection. Sparse approximation algorithm LASSO is used to reconstruct test video and reconstruction error is calculated for each of the dictionaries. To get another discriminative measure projection, the test vector is projected onto the atoms in the dictionary. Minimum reconstruction error and maximum projection give information regarding the action category of the test vector. With action bank as a feature vector, our best performance is 59.3% on UCF50 (benchmark is 57.9%), 97.7% on KTH (benchmark is 98.2%)and 23.63% on HMDB51 (benchmark is 26.9%).

The field of music and speech classification is quite mature with researchers having settled on the approximate best discriminative representation. In this regard, Zubair et al. showed the use of sparse coefficients along with SVM to classify audio signals as music or speech to get a near-perfect classification. In the proposed method, we go one step further, instead of using the sparse coefficients with another classifier they are directly used in a dictionary which is learned using on-line dictionary learning for music-speech classification. This approach removes the redundancy of using a separate classifier but also produces complete discrimination of music and speech on the GTZAN music/speech dataset. Moreover, instead of the high-dimensional feature vector space which inherently leads to high computation time and complicated decision boundary calculation on the part of SVM, the restricted dictionary size with limited computation serves the same purpose.

In this paper, an approach for music genre classification based on sparse representation using MARSYAS features is proposed. The MARSYAS feature descriptor consisting of timbral texture, pitch and beat related features is used for the classification of music genre. On-line Dictionary Learning (ODL) is used to achieve sparse representation of the features for developing dictionaries for each musical genre. We demonstrate the efficacy of the proposed framework on the Latin Music Database (LMD) consisting of over 3000 tracks spanning 10 genres namely Axé, Bachata, Bolero, Forró, Gaúcha, Merengue, Pagode, Salsa, Sertaneja and Tango.

Automatic medical image classification refers to assigning an medical image into a class, among a number of image categories. Due to computational complexity, it is an important task in the content based image retrieval (CBIR). In this paper, we propose a method for classification and retrieval of medical images using multi-feature extraction method. Here, edge and patch based methods are used to extract feature vectors. We demonstrate that these features provide discriminative information useful for classification of medical images by considering eight categories of images, namely, hand, skull, chest, mammogram, chest side view, skull side view, knee and neck. Using multi-feature extraction method improves the accuracy of CBIR. In this paper, three different types of similarity measures are used for comparing the query image with database images. The experimental results suggest that the proposed method has the ability to retrieve relevant images for a given input query image and provides good retrieval performance than compared with the single feature extraction method.

There are many clustering methods available and each of them may give a different grouping of datasets. It is proven that hybrid clustering algorithms give efficient results over the other algorithms. In this paper, we propose an efficient hybrid clustering algorithm by combining the features of leader's method which is an incremental clustering method and complete linkage algorithm which is a hierarchical clustering procedure. It is most common to find the dissimilarity between two clusters as the distance between their centorids or the distance between two closest (or farthest) data points. However, these measures may not give efficient clustering results in all cases. So, we propose a new similarity measure, known as cohesion to find the intercluster distance. By using this measure of cohesion, a two level clustering algorithm is proposed, which runs in linear time to the size of input data set. We demonstrate the effectiveness of the clustering procedure by using the leader's algorithm and cohesion similarity measure. The proposed method works in two steps: In the first step, the features of incremental and hierarchical clustering methods are combined to partition the input data set into several smaller subclusters. In the second step, subclusters are merged continuously based on cohesion similarity measure. We demonstrate the effectiveness of this framework for the web mining applications.