Anis Ben Ammar

Bio: Anis Ben Ammar is an academic researcher from University of Sfax. The author has contributed to research in topics: Image retrieval & Semantic similarity. The author has an hindex of 10, co-authored 52 publications receiving 350 citations.

A fuzzy ontology: based framework for reasoning in visual video content analysis and indexing

Abstract: Multimedia indexing systems based on semantic concept detectors are incomplete in the semantic sense. We can improve the effectiveness of these systems by using knowledge-based approaches which utilize semantic knowledge. In this paper, we propose a novel and efficient approach to enhance semantic concept detection in multimedia content, by exploiting contextual information about concepts from visual modality. First, a semantic knowledge is extracted via a contextual annotation framework. Second, a Fuzzy ontology is proposed to represent the fuzzy relationships (roles and rules) among every context and its semantic concepts. We use an abduction engine based on βeta function as a membership function for fuzzy rules. Third, a deduction engine is used to handle richer results in our video indexing system by running the proposed fuzzy ontology. Experiments on TRECVID 2010 benchmark have been performed to evaluate the performance of this approach. The obtained results show consistent improvement in semantic concepts detection, when a context space is used, and a good degree of indexing effectiveness as compared to existing approaches.

Effective Diversification for Ambiguous Queries in Social Image Retrieval

Abstract: Recent years have witnessed a great popularity of social photos sharing websites, which host a tremendous volume of digital images accompanied by their associated tags. Thus, extensive research efforts have been dedicated to tag-based social image search which enables users to formulate their queries using tags. However, tag queries are often ambiguous and typically short. Search results diversification approach is the common solution which aims to increase the number of satisfied users using only a single results set that cover the maximum of query aspects. However, not all queries are uniformly ambiguous and hence different diversification strategies might be suggested. In such context, we propose a new ranking process which dynamically predicts an effective trade-off between the relevance and diversity based results ranking according to the ambiguity level of a given query. Thorough experiments using 12 ambiguous queries over the NUS-WIDE dataset show the effectiveness of our approach over classical uniform diversification approaches.

Adaptive diversification for tag-based social image retrieval

Abstract: With the great popularity of social photos sharing websites, a tremendous volume of digital images is hosted together with their associated tags. Thus, extensive research efforts have been dedicated to tag-based social image search which enables users to formulate their queries using tags. However, tag queries are often ambiguous and typically short. Diversifying search results is a common solution in the absence of further knowledge about the user’s intention. Such approach aims to retrieve relevant images covering as much of the diverse meanings the query may have. However, not all queries are uniformly ambiguous and hence different diversification strategies might be suggested. In such a context, two new processes are jointly investigated at query pre-processing and post-processing levels. On the one hand, we propose a multi-view concept-based query expansion process, using a predefined list of semantic concepts, which aims to weight concepts from different views or contexts, aggregate the obtained weights and select the most representative ones using a dynamic threshold. On the other hand, we propose a new ranking process called “adaptive diverse relevance ranking” which automatically predicts an effective trade-off between relevance scores and diversity scores according to the query ambiguity level. Thorough experiments using 12 ambiguous queries over the NUS-WIDE dataset show the effectiveness of our approach versus classical uniform diversification approaches.

Improving image search effectiveness by integrating contextual information

Abstract: Recently, image retrieval approaches shift to context-based reasoning. Context-based approaches proved their efficiency to improve retrieval process. In fact, conventional image search engines are often not able to satisfy the user's intent as they provide noisy or/and redundant results. In addition, when a query is ambiguous, such systems can hardly distinguish different meanings for one query and therefore, they fail to show images with different contexts. A good system should provide, at top-k results, images which are the most relevant and diverse to guarantee user's satisfaction. Our objective is to improve the retrieval process performance by harnessing the contextual information to measure the relevance score and diversity score. The proposed approach implies the relevance-based ranking where a random walk with restart offers a refining step, the diversity-based ranking and the combination. Our approach was evaluated in the context of ImageCLEF1 benchmark. Obtained results are promising especially for diversity-based ranking.

REGIM at TRECVID2008: High-level Features Extraction and Video Search.

Abstract: A telescoping, outdoor, lighted, Christmas tree exhibit, which at night will simulate a lighted Christmas tree. The Christmas tree exhibit comprises a plurality of concentric hoops, populated with Christmas lights. The hoops are each of decreasing diameter that are sequentially interconnected by a plurality of concentric non stretching, flexible plastic bands, each hoop decreasing in size at the same rate as the hoop before it, with the largest-diameter hoop at the bottom of the Christmas tree exhibit and the smallest-diameter hoop at the top of the Christmas tree exhibit, when fully extended, a conical, tree-shaped appearance is obtained. Preceding the uppermost hoop is a top disk that is connected to the top end of the non stretching, flexible plastic bands. The top disk is supported by an adjustable, collapsible center pole. The bottom end of the non stretching, flexible plastic bands are anchored at the bottom of Christmas tree exhibit, so that when the non stretching, flexible plastic bands are fully extended and pulled taut, the hoops of the tree are completely suspended off the ground and the Christmas tree is standing. The tree is light weight and wind resistant. When disassembled, the hoops rest one inside the other, not touching each other, allowing the lights to stay attached to the hoops while in storage and preventing the breakage of bulbs and entangling of lights, resulting in a quick pop-up installation, and quick disassembly with minimal storage space required.

The Self-Organizing Map

Abstract: An overview of the self-organizing map algorithm, on which the papers in this issue are based, is presented in this article.

A survey on indexing techniques for big data: taxonomy and performance evaluation

Abstract: The explosive growth in volume, velocity, and diversity of data produced by mobile devices and cloud applications has contributed to the abundance of data or `big data.' Available solutions for efficient data storage and management cannot fulfill the needs of such heterogeneous data where the amount of data is continuously increasing. For efficient retrieval and management, existing indexing solutions become inefficient with the rapidly growing index size and seek time and an optimized index scheme is required for big data. Regarding real-world applications, the indexing issue with big data in cloud computing is widespread in healthcare, enterprises, scientific experiments, and social networks. To date, diverse soft computing, machine learning, and other techniques in terms of artificial intelligence have been utilized to satisfy the indexing requirements, yet in the literature, there is no reported state-of-the-art survey investigating the performance and consequences of techniques for solving indexing in big data issues as they enter cloud computing. The objective of this paper is to investigate and examine the existing indexing techniques for big data. Taxonomy of indexing techniques is developed to provide insight to enable researchers understand and select a technique as a basis to design an indexing mechanism with reduced time and space consumption for BD-MCC. In this study, 48 indexing techniques have been studied and compared based on 60 articles related to the topic. The indexing techniques' performance is analyzed based on their characteristics and big data indexing requirements. The main contribution of this study is taxonomy of categorized indexing techniques based on their method. The categories are non-artificial intelligence, artificial intelligence, and collaborative artificial intelligence indexing methods. In addition, the significance of different procedures and performance is analyzed, besides limitations of each technique. In conclusion, several key future research topics with potential to accelerate the progress and deployment of artificial intelligence-based cooperative indexing in BD-MCC are elaborated on.

A review of deep learning with special emphasis on architectures, applications and recent trends

Abstract: Deep learning (DL) has solved a problem that a few years ago was thought to be intractable — the automatic recognition of patterns in spatial and temporal data with an accuracy superior to that of humans. It has solved problems beyond the realm of traditional, hand-crafted machine learning algorithms and captured the imagination of practitioners who are inundated with all types of data. As public awareness of the efficacy of DL increases so does the desire to make use of it. But even for highly trained professionals it can be daunting to approach the rapidly increasing body of knowledge in the field. Where does one start? How does one determine if a particular DL model is applicable to their problem? How does one train and deploy them? With these questions in mind, we present an overview of some of the key DL architectures. We also discuss some new automatic architecture optimization protocols that use multi-agent approaches. Further, since guaranteeing system uptime is critical to many applications, a section dwells on using DL for fault detection and mitigation. This is followed by an exploratory survey of several areas where DL emerged as a game-changer: fraud detection in financial applications, financial time-series forecasting, predictive and prescriptive analytics, medical image processing, power systems research and recommender systems. The thrust of this review is to outline emerging applications of DL and provide a reference to researchers seeking to use DL in their work for pattern recognition with unparalleled learning capacity and the ability to scale with data.