Mouna Dammak

Histogram of dense subgraphs for image representation

Modelling spatial information of local features is known to improve performance in image categorisation. Compared with simple pairwise features and visual phrases, graphs can capture the structural organisation of local features more adequately. Besides, a dense regular grid can guarantee a more reliable representation than the interest points and give better results for image classification. In this study, the authors introduced a bag of dense local graphs approach that combines the performance of bag of visual words expressing the image classification process with the representational power of graphs. The images were represented with dense local graphs built upon dense scale-invariant feature transform descriptors. The graph-based substructure pattern mining algorithm was applied on the local graphs to discover the frequent local subgraphs, producing a bag of subgraphs representation. The results were reported from experiments conducted on four challenging benchmarks. The findings show that the proposed subgraph histogram improves the categorisation accuracy.

Laplacian Tensor sparse coding for image categorization

To generate the visual codebook, a step of quantization process is obligatory. Several works have proved the efficiency of sparse coding in feature quantization process of BoW based image representation. Furthermore, it is an important method which encodes the original signal in a sparse signal space. Yet, this method neglects the relationships among features. To reduce the impact of this issue, we suggest in this paper, a Laplacian Tensor sparse coding method, which will aim to profit from the relationship among the local features. Precisely, we propose to apply the similarity of tensor descriptors to create a Laplacian Tensor similarity matrix, which can better present in the same time the closeness of local features in the data space and the topological relationship among the spatially near local descriptors. Moreover, we integrate statistical analysis applied to the local features assigned to each visual word in the pooling step. Our experimental results prove that our method prevails or exceeds existing background results.

Flickr-based semantic context to refine automatic photo annotation

Automatic photo annotation task aims to describe the semantic content by detecting high level concepts in order to further facilitate concept based video retrieval. Most of existing approaches are based on independent semantic concept detectors without considering the contextual correlation between concepts. This drawback has its impact over the efficiency of such systems. Recently, harnessing contextual information to improve the effectiveness of concepts detection becomes a promising direction in such field. In this paper, we propose a new contextbased annotation refinement process. For this purpose, we define a new semantic measure called “Second Order Co-occurence Flickr context similarity” (SOCFCS) which aims to extract the semantic context correlation between two concepts by exploring Flickr resources (Flickr related-tags). Our measure is an extension of FCS measure by taking into consideration the FCS values of common Flickr related-tags of the two target concepts. Our proposed measure is applied to build a concept network which models the semantic context inter-relationships among concepts. A Random Walk with Restart process is performed over this network to refine the annotation results by exploring the contextual correlation among concepts. Experimental studies are conducted on ImageCLEF 2011 Collection containing 10000 images and 99 concepts. The results demonstrate the effectiveness of our proposed approach.

Extending Laplacian sparse coding by the incorporation of the image spatial context

Diverse studies have shown the efficiency of sparse coding in feature quantization. However, its major drawback is that it neglects the relationships among features. To reach the spatial context, we proposed in this paper, a novel sparse coding method called Extended Laplacian Sparse Coding. Two successive stages are required in this method. In the first stage, the sparse visual phrases based on Laplacian sparse coding are generated from the local regions in order to represent the geometric information in the image space. The second stage aims to incorporate the spatial relationships among local features in the image space into the objective function of the Laplacian sparse coding. It takes into account the similarity among local regions in the Laplacian sparse coding process. The matching between the local regions is based on the Hungarian method as well as the histogram intersection measure between sparse visual phrases already assigned to the local regions in the first stage. Furthermore, we suggested to improve the pooling step that succeeds the encoding step by introducing the discretized max pooling method that estimates the distribution of the responses of each local feature to the dictionary of basis vectors. Our experimental results prove that our method outperforms the existing background results.

Efficacy of Bacillus subtilis V26 as a biological control agent against Rhizoctonia solani on potato.

The aim of this study is to evaluate the efficacy of the strain Bacillus subtilis V26, a local isolate from the Tunisian soil, to control potato black scurf caused by Rhizoctonia solani. The in vitro antifungal activity of V26 significantly inhibited R. solani growth compared to the untreated control. Microscopic observations revealed that V26 caused considerable morphological deformations of the fungal hyphae such as vacuolation, protoplast leakage and mycelia crack. The most effective control was achieved when strain V26 was applied 24h prior to inoculation (protective activity) in potato slices. The antagonistic bacterium V26 induced significant suppression of root canker and black scurf tuber colonization compared to untreated controls with a decrease in incidence disease of 63% and 81%, respectively, and promoted plant growth under greenhouse conditions on potato plants. Therefore, B. subtilis V26 has a great potential to be commercialized as a biocontrol agent against R. solani on potato crops.

Extended Laplacian Sparse Coding for Image Categorization

In image classification task, several recent works show that sparse representation plays a basic role in dictionary learning. However, this approach neglects the spatial relationships in the image space during dictionary learning. However, this approach neglects the neighboring relationship in dictionary learning. To alleviate the impact of this problem, we propose a novel dictionary learning based on Laplacian sparse coding method that profits from the neighboring relationship among the local features. For that purpose, we incorporate the matching between local regions in the Laplacian sparse coding formula. Moreover, we integrate statistical analysis of the distribution of the responses of each local feature to the dictionary basis in the final image representation. Our experimental results prove that our method performs existing background results based on sparse representation.

Modelling Tetraselmis sp. growth-kinetics and optimizing bioactive-compound production through environmental conditions

The aim of this study is to predict Tetraselmis cells growth-kinetic and to induce the synthesis of bioactive compounds (chlorophylls, carotenoids and starch) with high potential for biotechnological applications. Using the statistical criteria, the Baranyi-Roberts model has been selected to estimate the microalgae growth-kinetic values. The simultaneous effects of salinity, light intensity and pH of culture medium were investigated to maximize the production of total chlorophylls, carotenoids and starch. The optimal culture conditions for the production of these compounds were found using Box-Behnken Design. Results have shown that total chlorophyll and carotenoids were attained 21.6mg·g-1DW and 0.042mg·g-1DW, respectively. In addition, the highest starch content of 0.624g·g-1DW has been obtained at neutral pH with high irradiance (182μmolphotonsm-2 s-1) and low salinity (20). A highly correlation (R2 = 0.884) has been found between the gravimetric and flow cytometric measurements of chlorophyll content.

Cyanobacteria as source of marine bioactive compounds: Molecular specific detection based on Δ9 desaturase gene

The blue-green microalga, Arthrospira sp., isolated from the sea of Kssour Essef in Mahdia (Tunisia), was purified and then identified both morphologically and genetically based on 16S rRNA gene sequence. Following physicochemical analysis, the prokaryotic microalga tested represented a competitive source of pigments and showed a considerable rate in protein (64%) which was confirmed by FTIR measurement. The lipid content (4%) was quantified by the gravimetric method and the intracellular lipid bodies were detected with the Nile red staining. Using gas chromatography coupled with flame ionization detector, the fatty acid profile revealed the presence of 27.4% and 32.88% of monounsaturated fatty acids (MUFAs) and polyunsaturated fatty acids (PUFAs), respectively. Given the richness of the isolated microalga in unsaturated fatty acids, we have developed a SYBR Green real time PCR method for the specific identification of Arthrospira sp. Δ9 desaturase gene. This current method will be of great value for carrying out high-throughput studies like cloning, heterologous expression and structure-function relationship analysis.

A novel dictionary learning algorithm for image representation

Sparse coding has proved its efficiency in the image classification task. However, its major drawback is the discarding of the spatial context information that can be extracted from the image. Therefore, we propose in this work a novel sparse coding method called Laplacian sparse coding based on the integration of topological information in the encoding process. This is achieved by embedding the similarities between local region visual phrases into the objective function of the classical Laplacian sparse coding. Experimental results made on several datasets prove the efficiency of the proposed method.

Optimization of lipids’ ultrasonic extraction and production from Chlorella sp. using response-surface methodology

BackgroundThree steps are very important in order to produce microalgal lipids: (1) controlling microalgae cultivation via experimental and modeling investigations, (2) optimizing culture conditions to maximize lipids production and to determine the fatty acid profile the most appropriate for biodiesel synthesis, and (3) optimizing the extraction of the lipids accumulated in the microalgal cells.MethodsFirstly, three kinetics models, namely logistic, logistic-with-lag and modified Gompertz, were tested to fit the experimental kinetics of the Chlorella sp. microalga culture established on standard conditions. Secondly, the response-surface methodology was used for two optimizations in this study. The first optimization was established for lipids production from Chlorella sp. culture under different culture conditions. In fact, different levels of nitrate concentrations, salinities and light intensities were applied to the culture medium in order to study their influences on lipids production and determine their fatty acid profile. The second optimization was concerned with the lipids extraction factors: ultrasonic’s time and temperature, and chloroform-methanol solvent ratio.ResultsAll models (logistic, logistic-with-lag and modified Gompertz) applied for the experimental kinetics of Chlorella sp. show a very interesting fitting quality. The logistic model was chosen to describe the Chlorella sp. kinetics, since it yielded the most important statistical criteria: coefficient of determination of the order of 94.36%; adjusted coefficient of determination equal to 93.79% and root mean square error reaching 3.685 cells · ml− 1.Nitrate concentration and the two interactions involving the light intensity (Nitrate concentration × light intensity, and salinities × light intensity) showed a very significant influence on lipids production in the first optimization (p < 0.05). Yet, only the quadratic term of chloroform-methanol solvent ratio showed a significant influence on lipids extraction relative to the second step of optimization (p < 0.05).The two most abundant fatty acid methyl esters (≈72%) derived from the Chlorella sp. microalga cultured in the determined optimal conditions are: palmitic acid (C16:0) and oleic acid (C18:1) with the corresponding yields of 51.69% and 20.55% of total fatty acids, respectively.ConclusionsOnly the nitrate deficiency and the high intensity of light can influence the microalgal lipids production. The corresponding fatty acid methyl esters composition is very suitable for biodiesel production. Lipids extraction is efficient only over long periods of time when using a solvent with a 2/1 chloroform/methanol ratio.