Bogdan Cramariuc

An efficient watershed segmentation algorithm suitable for parallel implementation

An important aspect of designing a parallel algorithm is exploitation of the data locality for minimization of the communication overhead. We propose a reformulation of a global image operation called the watershed transformation. The method is one of the various approaches for image segmentation and works by labeling connected components. Both serial and parallel programming models are presented and evaluated when running on SUN and DEC Alpha AXP workstations, and a Cray T3D, respectively.

Parallel watershed transformation algorithms for image segmentation

Abstract The watershed transformation is a mid-level operation used in morphological image segmentation. Techniques applied on large images, which must often complete fast, are usually computationally expensive and complex entailing efficient parallel algorithms. Two distributed approaches of the watershed transformation are introduced in this paper. The algorithms survey in a Single Program Multiple Data (SPMD) model both local and global connectivity properties of the morphological gradient of a gray-scale image to label connected components. The sequentiality of the serial algorithm is broken in the parallel versions by exploiting the ordering relation between two neighboring pixels successively incorporated in the same region. Thus, a path is traced, for every unlabeled pixel, down to its region of inclusion (whose label is then propagated backwards); in the second algorithm, regions grow independently around their seeds. In both cases only pixels which satisfy the ordering relation are incorporated in any region. This way, not only different regions are explored in a parallel fashion, but also different parts of the same region, when the latter extends to neighboring subdomains, are treated likewise. Running time and relative speedup evaluated on a Cray T3D parallel computer are used to appreciate the performance of both algorithms.

An ordinal co-occurrence matrix framework for texture retrieval

We present a novel ordinal co-occurrence matrix framework for the purpose of content-based texture retrieval. Several particularizations of the framework will be derived and tested for retrieval purposes. Features obtained using the framework represent the occurrence frequency of certain ordinal relationships at different distances and orientations. In the ordinal co-occurrence matrix framework, the actual pixel values do not affect the features, instead, the ordinal relationships between the pixels are taken into account. Therefore, the derived features are invariant to monotonic gray-level changes in the pixel values and can thus be applied to textures which may be obtained, for example, under different illumination conditions. Described ordinal co-occurrence matrix approaches are tested and compared against other well-known ordinal and nonordinal methods.

Texture similarity evaluation using ordinal co-occurrence

Co-occurrence matrices have been successfully used in texture analysis. However, due to noise and monotonic shifts in gray levels, traditional co-occurrence analysis may lead to erroneous results. Using the order of the gray values instead of the gray values themselves is shown to improve the retrieval accuracy. Ordinal measures have been used for many image processing tasks in the literature. In this paper, we propose a novel combination of ordinal measures and co-occurrence matrices using local pixel pair comparisons. Features constructed in this paper represent the occurrence frequency of certain ordinal relationships at different distances and orientations. The proposed method gives encouraging results when comparing its retrieval performance to that of the traditional gray level co-occurrence matrices.

Relevance feedback for shape query refinement

In this paper we propose to incorporate a feedback loop, into the ordinal correlation framework and apply it to shape-based image retrieval. The users feedback on the relevance of the retrieval results is used to tune the weights of the similarity measure. Statistics from the features of both relevant and irrelevant items are used to estimate the weights. Moreover, the information accumulated from previous retrieval iterations is used in the weights estimation. A simple measure of the discrimination power is proposed and used to show that the relevance feedback increases the capability of the ordinal correlation scheme to discriminate between relevant and irrelevant objects.

Block-based ordinal co-occurrence matrices for texture similarity evaluation

In this paper we introduce a block-based approach for ordinal co-occurrence matrices aimed at improving robustness of the basic ordinal co-occurrence. Earlier, we have introduced two approaches for building ordinal co-occurrence matrices. One considers only the center pixel of a moving window as a seed point, compares it to its anti-causal neighbors and saves the occurrences of ordinal relations between pixels in the form of co-occurrence matrices. However, in that approach problems occur especially when considering textures with slightly varying gray levels in relatively large areas. The other method improves the robustness of the earlier method by considering also other pixels than the center pixel in the thresholded window as seed points. Main drawback of that method is the increased computational complexity. In order to avoid that, a block-based approach for building the ordinal co-occurrence matrices is introduced in this paper. Retrieval accuracy of the proposed method has shown to produce better results than existing techniques.

User requirements in the context of future location based services as seen from a survey among Romanian students

Location Based Services (LBS) are nowadays recognized as having one of the most fast-paced developments in the wireless field. The underlying location technology is still mainly based on the Global Navigation Satellite Systems (GNSS), but many complementary technologies, such as cellular-based location, WLAN-based location and other Signal of Opportunity-based location are rapidly trying to close the remaining technology gaps. There is a myriad of potential future applications for LBS, but their perceived value for the user is still not well-studied or understood. The goal of our paper is to better define the user requirements in terms of mobile device features and LBS applications, based on a survey among 47 students at a top technical university in Bucharest, Romania.

Multidimensional texture characterization using moments

In this paper a moment based texture characterization technique is presented. The texture features are obtained directly from the gray-level image by computing the moments of the image in local regions. A texture is represented as a constellation of clusters in the multidimensional feature space. The computation of certain invariants is also mentioned.

A Framework for Activity Recognition Through Deep Learning and Abnormality Detection in Daily Activities

Activity recognition plays a key role in providing activity assistance and care for users in intelligent homes. This paper presents a two layer of convolutional neural networks to perform human action recognition using images provided by multiple cameras. We consider one PTZ camera and multiple Kinects in order to offer continuity over the users movement. The drawbacks of using only one type of sensor is minimized. For example, field of view provided by Kinect sensor is not wide enough to cover the entire room. Also, the PTZ camera is not able to detect and track a person in case of different situations, such as the person is sitting or it is under the camera. Also the system will identify abnormalities that can appear in sequences of performed daily activities. The system is tested in Ambient Intelligence Laboratory (AmI-Lab) at the University Politehnica of Bucharest.

Mobile@Old - an assistive platform for maintaining a healthy lifestyle for elderly people

Ambient Assistive Living (AAL) applications allow elderly people to maintain a healthy life style and live longer in their homes. In this paper we describe a platform that combine physical exercises, health monitoring with a reminder component implemented as a multimodal interface adapted to elderly needs for maintaining a healthy lifestyle for elderly people.