Filters








38 Hits in 1.4 sec

Magnification Generalization for Histopathology Image Embedding [article]

Milad Sikaroudi, Benyamin Ghojogh, Fakhri Karray, Mark Crowley, H.R. Tizhoosh
2021 arXiv   pre-print
Histopathology image embedding is an active research area in computer vision. Most of the embedding models exclusively concentrate on a specific magnification level. However, a useful task in histopathology embedding is to train an embedding space regardless of the magnification level. Two main approaches for tackling this goal are domain adaptation and domain generalization, where the target magnification levels may or may not be introduced to the model in training, respectively. Although
more » ... fication adaptation is a well-studied topic in the literature, this paper, to the best of our knowledge, is the first work on magnification generalization for histopathology image embedding. We use an episodic trainable domain generalization technique for magnification generalization, namely Model Agnostic Learning of Semantic Features (MASF), which works based on the Model Agnostic Meta-Learning (MAML) concept. Our experimental results on a breast cancer histopathology dataset with four different magnification levels show the proposed method's effectiveness for magnification generalization.
arXiv:2101.07757v1 fatcat:d3z7y2fmtbburanhu32jz32b5u

Weighted Fisher Discriminant Analysis in the Input and Feature Spaces [article]

Benyamin Ghojogh, Milad Sikaroudi, H.R. Tizhoosh, Fakhri Karray, Mark Crowley
2020 arXiv   pre-print
Fisher Discriminant Analysis (FDA) is a subspace learning method which minimizes and maximizes the intra- and inter-class scatters of data, respectively. Although, in FDA, all the pairs of classes are treated the same way, some classes are closer than the others. Weighted FDA assigns weights to the pairs of classes to address this shortcoming of FDA. In this paper, we propose a cosine-weighted FDA as well as an automatically weighted FDA in which weights are found automatically. We also propose
more » ... a weighted FDA in the feature space to establish a weighted kernel FDA for both existing and newly proposed weights. Our experiments on the ORL face recognition dataset show the effectiveness of the proposed weighting schemes.
arXiv:2004.01857v1 fatcat:uqsaqubyxbbyxloc3mepwyuteq

Fisher Discriminant Triplet and Contrastive Losses for Training Siamese Networks [article]

Benyamin Ghojogh, Milad Sikaroudi, Sobhan Shafiei, H.R. Tizhoosh, Fakhri Karray, Mark Crowley
2020 arXiv   pre-print
Siamese neural network is a very powerful architecture for both feature extraction and metric learning. It usually consists of several networks that share weights. The Siamese concept is topology-agnostic and can use any neural network as its backbone. The two most popular loss functions for training these networks are the triplet and contrastive loss functions. In this paper, we propose two novel loss functions, named Fisher Discriminant Triplet (FDT) and Fisher Discriminant Contrastive (FDC).
more » ... The former uses anchor-neighbor-distant triplets while the latter utilizes pairs of anchor-neighbor and anchor-distant samples. The FDT and FDC loss functions are designed based on the statistical formulation of the Fisher Discriminant Analysis (FDA), which is a linear subspace learning method. Our experiments on the MNIST and two challenging and publicly available histopathology datasets show the effectiveness of the proposed loss functions.
arXiv:2004.04674v1 fatcat:hc6ye3nbl5g6jgmn25atep3wuu

Query Classification using Wikipedia's Category Graph

Milad AlemZadeh, Richard Khoury, Fakhri Karray
2012 Journal of Emerging Technologies in Web Intelligence  
Wikipedia's category graph is a network of 300,000 interconnected category labels, and can be a powerful resource for many classification tasks. However, its size and the lack of order can make it difficult to navigate. In this paper, we present a new algorithm to efficiently exploit this graph and accurately rank classification labels given user-specified keywords. We highlight multiple possible variations of this algorithm, and study the impact of these variations on the classification
more » ... in order to determine the optimal way to exploit the category graph. We implement our algorithm as the core of a query classification system and demonstrate its reliability using the KDD CUP 2005 and TREC 2007 competitions as benchmarks.
doi:10.4304/jetwi.4.3.207-220 fatcat:pou4d5tnt5bfdhhtgratruyw3m

Association between Dichlorodiphenyldichloroethylene in the Serum and Adipose Tissue with Type 2 Diabetes: A Systematic Review and Meta-Analysis

Yadolah Fakhri, Amir Sheikhmohammadi, Abotaleb Bay, Yahya Zandsalimi, Bigard Moradi, Moayed Avazpour, Milad Moloudizargari, Milad Adel, Sayed Mohsen Mohseni, Hassan Keramati
2016 Global Journal of Health Science  
<p>The number of diabetics is growing every day. In addition to the main risk factors of type 2 diabetes (obesity, age and etc.) other environmental risk factors such as persistent organic pollutants are also considered. Dichlorodiphenyldichloroethylene is one of the persistent organic pollutants which are produced by the pesticide DDT metabolized and its effects of type 2 diabetes are taken into consideration of many investigators. Inconsistency in the results caused to try to achieve a
more » ... d assessment of the effect of dichlorodiphenyldichloroethylene on type 2 diabetes by systematic review and meta-analysis. In this review study after a systematic review, finally 6 prospective and 7 cross-sectional studies were obtained. After approving the heterogeneity of the studies based on random effect model, it was calculate the mean of odds ratio. Meta-analysis of the pooled random effects of Dichlorodiphenyldichloroethylene was done with type 2 diabetes in prospective and cross-sectional studies. In prospective studies, the highest and lowest odds ratio was 6.1(95%CI: 1.36-27.27) and 1.01(95%CI: 0.59-1.70) and cross-sectional studies was 5.5(95%CI: 1.2-25.10) and 0.61(95%CI: 0.32-1.15), respectively. In the all studies (sum of prospective and cross-sectional), in the random effect model odds ratio was 1.52 (95%CI: 1.26-1.84) and heterogeneity was 52.1% (ρ<sub>heterogeneity</sub>=0.009). Increasing concentration of Dichlorodiphenyldichloroethylene in the serum and adipose tissue increased the risk of type 2 diabetes significantly (P&lt;0.001). The results of this review study support the role of dichlorodiphenyldichloroethylene as an environmental risk factor for type 2 diabetes.</p>
doi:10.5539/gjhs.v9n2p43 fatcat:gekso2ghwfh7rhoaq5hthinniu

Batch-Incremental Triplet Sampling for Training Triplet Networks Using Bayesian Updating Theorem [article]

Milad Sikaroudi, Benyamin Ghojogh, Fakhri Karray, Mark Crowley, H.R. Tizhoosh
2020 arXiv   pre-print
Variants of Triplet networks are robust entities for learning a discriminative embedding subspace. There exist different triplet mining approaches for selecting the most suitable training triplets. Some of these mining methods rely on the extreme distances between instances, and some others make use of sampling. However, sampling from stochastic distributions of data rather than sampling merely from the existing embedding instances can provide more discriminative information. In this work, we
more » ... mple triplets from distributions of data rather than from existing instances. We consider a multivariate normal distribution for the embedding of each class. Using Bayesian updating and conjugate priors, we update the distributions of classes dynamically by receiving the new mini-batches of training data. The proposed triplet mining with Bayesian updating can be used with any triplet-based loss function, e.g., triplet-loss or Neighborhood Component Analysis (NCA) loss. Accordingly, Our triplet mining approaches are called Bayesian Updating Triplet (BUT) and Bayesian Updating NCA (BUNCA), depending on which loss function is being used. Experimental results on two public datasets, namely MNIST and histopathology colorectal cancer (CRC), substantiate the effectiveness of the proposed triplet mining method.
arXiv:2007.05610v2 fatcat:r6rvbmbkbbabndo6ljghclkcqq

Drifting Speed of Lagrangian Fronts and Oil Spill Dispersal at the Ocean Surface

Gina Fifani, Alberto Baudena, Milad Fakhri, Georges Baaklini, Yannice Faugère, Rosemary Morrow, Laurent Mortier, Francesco d'Ovidio
2021 Remote Sensing  
Due to its dire impacts on marine life, public health, and socio-economic services, oil spills require an immediate response. Effective action starts with good knowledge of the ocean dynamics and circulation, from which Lagrangian methods derive key information on the dispersal pathways present in the contaminated region. However, precise assessments of the capacity of Lagrangian methods in real contamination cases remain rare and limited to large slicks spanning several hundreds of km. Here we
more » ... address this knowledge gap and consider two medium-scale (tens of km wide) events of oil in contrasting conditions: an offshore case (East China Sea, 2018) and a recent near-coastal one (East Mediterranean, 2021). Our comparison between oil slicks and Lagrangian diagnostics derived from near-real-time velocity fields shows that the calculation of Lagrangian fronts is, in general, more robust to errors in the velocity fields and more informative on the dispersion pathways than the direct advection of a numerical tracer. The inclusion of the effect of wind is also found to be essential, being capable of suddenly breaking Lagrangian transport barriers. Finally, we show that a usually neglected Lagrangian quantity, the Lyapunov vector, can be exploited to predict the front drifting speed, and in turn, its future location over a few days, on the basis of near-real-time information alone. These results may be of special relevance in the context of next-generation altimetry missions that are expected to provide highly resolved and precise near-real-time velocity fields for both open ocean and coastal regions.
doi:10.3390/rs13224499 fatcat:qouci6h5hrgkhmpnjrj6pdg5he

Modelling surface currents in the Eastern Levantine Mediterranean using surface drifters and satellite altimetry

Leila Issa, Julien Brajard, Milad Fakhri, Daniel Hayes, Laurent Mortier, Pierre-Marie Poulain
2016 Ocean Modelling  
We present a new and fast method for blending altimetry and surface drifters data in the Eastern Levantine Mediterranean. The method is based on a variational assimilation approach for which the velocity is corrected by matching real drifters positions with those predicted by a simple advection model, while taking into account the wind effect. The velocity correction is done in a time-continuous fashion by assimilating at once a whole trajectory of drifters using a sliding time window. Except
more » ... r the wind component, the velocity is constrained to be divergence free. We show that with few drifters, our method improves the the estimation of velocity in two typical situations: an eddy between the Lebanese coast and Cyprus, and velocities along the Lebanese coast.
doi:10.1016/j.ocemod.2016.05.006 fatcat:4hnd7uhaojb3zhnwn72gtutj5u

Hesperetin-Loaded Solid Lipid Nanoparticles and Nanostructure Lipid Carriers for Food Fortification: Preparation, Characterization, and Modeling

Milad Fathi, Jaleh Varshosaz, Mohebbat Mohebbi, Fakhri Shahidi
2012 Food and Bioprocess Technology  
Solid lipid nanoparticles and nanostructure lipid carriers were used to entrap hesperetin and broaden confined knowledge of application of nanocarriers as the functional ingredients in food sectors. The produced nanocarriers using a high mechanical shear method were subjected to size and zeta potential analysis. The developed nanosize carriers had the encapsulation efficiency ranging from 39.90 to 63.08 %. Differential scanning calorimetry, X-ray diffraction, and Fourier transform infrared
more » ... roscopy were also employed to study thermal behavior, crystalline state, and chemical structure. The release behavior of hesperetin in simulated gastrointestinal conditions was investigated and kinetically modeled. The modeling results indicated that the release phenomenon is mostly governed by combination of Fickian and dissolution mechanisms. Stability of the nanocarriers, as analyzed for up to 30 days, at 6 and 25°C in aqueous suspension, showed no detectable hesperetin leakage. Cryoprotectant effect of different compounds (i.e., glucose, sorbitol, glycerin, lactose, and sucrose) was also examined. Finally, the potential capability of nanocarriers for food fortification was studied using milk as a model food. The fortified milk samples were subjected to sensory analysis and results betokened that the developed nanocarriers did not show any significant difference with blank milk sample and could well mask the bitter taste, after taste, and obviate poor solubility of hesperetin.
doi:10.1007/s11947-012-0845-2 fatcat:idksm6h3ovan5hag6wg6ufgz2e

Population biology of the common pandora, Pagellus erythrinus (Linnaeus, 1758) along the Lebanese coast, Eastern Mediterranean

Myriam Lteif, Sharif Jemaa, Rita Mouawad, Gaby Khalaf, Stefano Lelli, Milad Fakhri
2020 Egyptian Journal of Aquatic Research  
Pagellus erythrinus is a commercial fish in Lebanese marine waters. This species has been exploited by artisanal fisheries using gill or trammel nets, longlines and beach seines. Therefore, monthly biological P. erythrinus data have been collected since 2015 by the National Center for Marine Sciences -National Council for Scientific Research. Data from 2015, 2016, and 2017 were used to assess the growth and biology of this species along the Lebanese coast. A total of 1315 females and 446 males
more » ... ere collected and the sex ratio showed significant bias from 1:1. The weight-length relationship showed a negative allometry (b = 2.912 and R 2 = 0.97), and the length ± SD at which 50% of the individuals were sexually mature was 16.38 ± 0.16 cm for both sexes. The average gonado-somatic index and condition factor demonstrated that the spawning period of the common pandora population is spring and revealed the well-being of the population. Ó 2020 National Institute of Oceanography and Fisheries. Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
doi:10.1016/j.ejar.2020.01.002 fatcat:nnfqzoolpvhdremym45k6pmlhy

An Efficient Method for Tagging a Query with Category Labels Using Wikipedia towards Enhancing Search Engine Results

Milad Alemzadeh, Fakhri Karray
2010 2010 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology  
This paper intends to present a straightforward, extensive, and noise resistant method for efficiently tagging a web query, submitted to a search engine, with proper category labels. These labels are intended to represent the closest categories related to the query which can ultimately be used to enhance the results of any typical search engine by either restricting the results to matching categories or enriching the query itself. The presented method effectively rules out noise words within a
more » ... uery, forms the optimal keyword packs using a density function, and returns a set of category labels which represent the common topics of the given query using Wikipedia category hierarchy.
doi:10.1109/wi-iat.2010.267 dblp:conf/webi/AlemzadehK10 fatcat:orsp2wtnaneqddtp4ff5f5x4eu

Offline versus Online Triplet Mining based on Extreme Distances of Histopathology Patches [article]

Milad Sikaroudi, Benyamin Ghojogh, Amir Safarpoor, Fakhri Karray, Mark Crowley, H.R. Tizhoosh
2020 arXiv   pre-print
We analyze the effect of offline and online triplet mining for colorectal cancer (CRC) histopathology dataset containing 100,000 patches. We consider the extreme, i.e., farthest and nearest patches to a given anchor, both in online and offline mining. While many works focus solely on selecting the triplets online (batch-wise), we also study the effect of extreme distances and neighbor patches before training in an offline fashion. We analyze extreme cases' impacts in terms of embedding distance
more » ... for offline versus online mining, including easy positive, batch semi-hard, batch hard triplet mining, neighborhood component analysis loss, its proxy version, and distance weighted sampling. We also investigate online approaches based on extreme distance and comprehensively compare offline, and online mining performance based on the data patterns and explain offline mining as a tractable generalization of the online mining with large mini-batch size. As well, we discuss the relations of different colorectal tissue types in terms of extreme distances. We found that offline and online mining approaches have comparable performances for a specific architecture, such as ResNet-18 in this study. Moreover, we found the assorted case, including different extreme distances, is promising, especially in the online approach.
arXiv:2007.02200v2 fatcat:jx4orjssgzhhfatz4wevylj4d4

Carcinogenic and non-carcinogenic health risks of metal(oid)s in tap water from Ilam city, Iran

Yadolah Fakhri, Narottam Saha, Sahebeh Ghanbari, Milad Rasouli, Ali Miri, Moayed Avazpour, Aziz Rahimizadeh, Seyed-Mohammad Riahi, Mansour Ghaderpoori, Hassan Keramati, Bigard Moradi, Nazak Amanidaz (+1 others)
2018 Food and Chemical Toxicology  
Heavy metals have the properties of biological accumulation, toxicity and environmental stability, hence consumption of drinking water containing heavy metals can jeopardize human health. One of these heavy metals is cadmium that its long-term exposure causes kidney diseases, osteoporosis, cancer and cardiovascular disease. In this cross-sectional study which was conducted in Minab, 100 samples of tap water were collected from 10 regions during December and January of 2014. Cadmium
more » ... was measured by graphite flame atomic absorption spectrophotometry of AAS8000 model. Then, its non-carcinogenic risk was calculated through EPA equations. Range and mean concentration of Cadmium in December and January is ND-4.6 µg/l, 1.8±0.69 µg/l and ND-3.6 µg/l, 1.6±0.6 µg/l, respectively. So, the mean concentration of Cadmium is 1.7±0.64 µg/l. The mean of daily chronic intake and non-carcinogenic risk of Minab population is 0.00005 µg/kg-day and 0.095. The mean concentration of Cadmium tap drinking water is lower than WHO and EPA standard limits. Since the non-carcinogenic risk is lower than 1, thus it can be said that Minab city is in the safe area in terms of non-carcinogenic risk of Cadmium of the drinking water. But, Valiasr and Soleghanareashave the highest and lowest non-carcinogenic risk, respectively.
doi:10.1016/j.fct.2018.04.039 pmid:29684495 fatcat:eqcc3rzspfcapev3d2pdkj7m6q

Coastal Monitoring in the Context of Climate Change: Time-Series Efforts in Lebanon and Argentina

Abed Hassoun, National Council for Scientific Research, National Center for Marine Sciences, Jounieh, Lebanon, Rodrigo-Hernández Moresino, Elena Barbieri, Juan Cruz Carbajal, Augusto Crespi-Abril, Antonela De Cian, Lucía Epherra, Milad Fakhri, Abeer Ghanem, Houssein Jaber, Marie-Thérèse Kassab (+6 others)
2021 Oceanography  
Barbieri, Juan Cruz Carbajal, Augusto Crespi-Abril, Antonella De Cian, Lucía Epherra, Milad Fakhri, Abeer Ghanem, Houssein Jaber, Marie-Thérèse Kassab, Antonela Martelli, Anthony Ouba, Flavio Paparazzo  ... 
doi:10.5670/oceanog.2021.supplement.02-05 fatcat:2bpeg4vs3fffpimx7qtceflbrq

Modeling Surface Currents in the Eastern Levantine Mediterranean Modeling surface currents in the Eastern Levan5ne Mediterranean

Leila Issa, Julien Brajard, Milad Fakhri, Laurent Mortier, Pierre-Marie Poulain, Leila Issa, Julien Brajard, Milad Fakhri, Laurent Mortier, Pierre-Marie Poulain, Surface, Leila Issa (+4 others)
2015 Currents in the Eastern Levantine Mediterranean. EGU General Assembly Conference   unpublished
Fakhri (Lebanese CNRS), Laurent Mor5er (UPMC, LOCEAN), and Pierre--Marie Poulain (OGS, Trieste) Abstract Figure 1 . 1 Velocity field from al5metric data and 3 driQer trajectories (real and simulated  ...  Ũ r =r b +~ r J ( Ũ ) = 1 2 ( X i X m kr b i + r i ( Ũ ) r o i (m t)k 2 + k~ U k 2 B ) T L Contacts : leila.issa@lau.edu.lb , julien.brajard@locean--ipsl.upmc.fr (LAU), Julien Brajard (UPMC, LOCEAN), Milad  ... 
fatcat:g6jh65ezi5bktl5z7qquwk65rq
« Previous Showing results 1 — 15 out of 38 results