Andrei Stoian - Academia.edu (original) (raw)

Papers by Andrei Stoian

Coatings

High-entropy alloys (HEAs), also known as multicomponent or multi-principal element alloys (MPEAs... more High-entropy alloys (HEAs), also known as multicomponent or multi-principal element alloys (MPEAs), differ from traditional alloys, which are usually based only on one principal element, in that they are usually fabricated from five or more elements in large percentages related to each other, in the range of 5%–35%. Despite the usually outstanding characteristics of HEAs, based on a properly selected design, many such alloys are coated with advanced composites after their elaboration to further improve their qualities. In this study, 73Ti-20Zr-5Ta-2Ag samples were covered with chitosan and a mixture of chitosan, bioglass, and ZnO particles to improve the materials’ antibacterial properties. A variety of methods, including scanning electron microscopy, atomic force microscopy, and mechanical and electrochemical determinations, has permitted a quantified comparison between the coated and uncoated surfaces of this medium–high-entropy alloy. The materials’ properties were enhanced by th...

Lecture Notes in Computer Science, 2019

We present a new replay-based method of continual classification learning that we term "condition... more We present a new replay-based method of continual classification learning that we term "conditional replay" which generates samples and labels together by sampling from a distribution conditioned on the class. We compare conditional replay to another replay-based continual learning paradigm (which we term "marginal replay") that generates samples independently of their class and assigns labels in a separate step. The main improvement in conditional replay is that labels for generated samples need not be inferred, which reduces the margin for error in complex continual classification learning tasks. We demonstrate the effectiveness of this approach using novel and standard benchmarks constructed from MNIST and FashionMNIST data, and compare to the regularization-based elastic weight consolidation (EWC) method [17, 34].

The electrochemical stability of CoCr based alloys is due to the passive oxide film grown on thei... more The electrochemical stability of CoCr based alloys is due to the passive oxide film grown on their surface, but it is to notice the aspect of ion release in their working time in this paper a new CoCrMo dental alloy alloy with small amount of niobium and zirconium (6% Nb and 0.8% Zr respectively) is modified by incorporation Ag on its surface. This paper aims to compare the ion release from CoCrMoNbZr in Tani Zuchi artificial saliva at various periods of time before and after Ag nanoparticles incorporation by pulselectrodeposition. The surface analysis reveled that Ag are nanoparticles with 100 nm average diameter. The ion release determination with inductively plasma mass spectrometer (ICP-MS established that after Ag incorporation the surface is more protective and Zr release decreased significantly. Electrochemical tests in artificial saliva Tani Zuchi sustain the ICP-MS determination which indicate better performance of alloy CoCr MoNbZr after Ag incorporation.

Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, 2018

In video surveillance, pedestrian attributes such as gender, clothing or hair types are useful cu... more In video surveillance, pedestrian attributes such as gender, clothing or hair types are useful cues to identify people. The main challenge in pedestrian attribute recognition is the large variation of visual appearance and location of attributes due to different poses and camera views. In this paper, we propose a neural network combining high-level learnt Convolutional Neural Network (CNN) features and low-level handcrafted features to address the problem of highly varying viewpoints. We first extract low-level robust Local Maximal Occurrence (LOMO) features and learn a body part-specific CNN to model attribute patterns related to different body parts. For small datasets which have few data, we propose a new learning strategy, where the CNN is pre-trained in a triplet structure on a person re-identification task and then fine-tuned on attribute recognition. Finally, we fuse the two feature representations to recognise pedestrian attributes. Our approach achieves state-of-the-art results on three public pedestrian attribute datasets.

Plastic and Reconstructive Surgery - Global Open, 2018

surgical technical difficulties, less complications and long term rates of survival.

Remote Sensing, 2021

Automatic ship detection provides an essential function towards maritime domain awareness for sec... more Automatic ship detection provides an essential function towards maritime domain awareness for security or economic monitoring purposes. This work presents an approach for training a deep learning ship detector in Sentinel-2 multi-spectral images with few labeled examples. We design a network architecture for detecting ships with a backbone that can be pre-trained separately. By using self supervised learning, an emerging unsupervised training procedure, we learn good features on Sentinel-2 images, without requiring labeling, to initialize our network’s backbone. The full network is then fine-tuned to learn to detect ships in challenging settings. We evaluated this approach versus pre-training on ImageNet and versus a classical image processing pipeline. We examined the impact of variations in the self-supervised learning step and we show that in the few-shot learning setting self-supervised pre-training achieves better results than ImageNet pre-training. When enough training data ar...

2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, 2021

We introduce in this paper a novel active learning algorithm for satellite image change detection... more We introduce in this paper a novel active learning algorithm for satellite image change detection. The proposed solution is interactive and based on a question & answer model, which asks an oracle (annotator) the most informative questions about the relevance of sampled satellite image pairs, and according to the oracle's responses, updates a decision function iteratively. We investigate a novel framework which models the probability that samples are relevant; this probability is obtained by minimizing an objective function capturing representativity, diversity and ambiguity. Only data with a high probability according to these criteria are selected and displayed to the oracle for further annotation. Extensive experiments on the task of satellite image change detection after natural hazards (namely tornadoes) show the relevance of the proposed method against the related work.

Materials and Corrosion, 2018

The paper studies the electrochemical stability in physiological serum and cell response of nanos... more The paper studies the electrochemical stability in physiological serum and cell response of nanostrucures obtained on Zr by anodizing in electrolytes containing fluorinated ethylene glycol and various amounts of K 2 CO 3. SEM analysis revealed that the porous nanostructure formation is greatly influenced by the various K 2 CO 3 concentrations in the electrolyte by tipping the balance between oxide formation and dissolution during the anodizing process. The porosities of the samples were evaluated from polarization resistances recorded in physiological serum. Chronoamperometric measurements were employed to establish a range of compactness for the samples. It was found that the amount of K 2 CO 3 from the electrolyte is directly responsible in obtaining oxide strata more resistant to corrosion. All tests performed on the samples such as corrosion, hydrophylicity and cell viability with MC3T3-E1 pre-osteoblasts cells from mouse bones revealed that the best coating for future research regarding nanostructured Zr implants was the one obtained from the electrolyte containing the highest amount of K 2 CO 3 .

Information Fusion, 2020

Continual learning (CL) is a particular machine learning paradigm where the data distribution and... more Continual learning (CL) is a particular machine learning paradigm where the data distribution and learning objective change through time, or where all the training data and objective criteria are never available at once. The evolution of the learning process is modeled by a sequence of learning experiences where the goal is to be able to learn new skills all along the sequence without forgetting what has been previously learned. CL can be seen as an online learning where knowledge fusion needs to take place in order to learn from streams of data presented sequentially in time. Continual learning also aims at the same time at optimizing the memory, the computation power and the speed during the learning process. An important challenge for machine learning is not necessarily finding solutions that work in the real world but rather finding stable algorithms that can learn in real world. Hence, the ideal approach would be tackling the real world in a embodied platform: an autonomous agent. Continual learning would then be effective in an autonomous agent or robot, which would learn autonomously through time about the external world, and incrementally develop a set of complex skills and knowledge. Robotic agents have to learn to adapt and interact with their environment using a continuous stream of observations. Some recent approaches aim at tackling continual learning for robotics, but most recent papers on continual learning only experiment approaches in simulation or with static datasets. Unfortunately, the evaluation of those algorithms does not provide insights on whether their solutions may help continual learning in the context of robotics. This paper aims at reviewing the existing state of the art of continual learning, summarizing existing benchmarks and metrics, and proposing a framework for presenting and evaluating both robotics and non robotics approaches in a way that makes transfer between both fields easier. We put light on continual learning in the context of robotics to create connections between fields and normalize approaches.

Journal of Clinical Oncology, 2018

e13572Background: Breast cancer (BC) is the most common malignancy in women with 6.6% of cases di... more e13572Background: Breast cancer (BC) is the most common malignancy in women with 6.6% of cases diagnosed in young women below the age of 40. Although there are significant similarities in demographic and clinico-pathological characteristic between geographical areas, the striking difference is in peak age for young populations. Objetive:Comparison of two young breast cancer(YBC) patients (pts) cohorts candidates for BRCA gene testing of two Developing Countries Methods: In this retrospective study, we analyzed 2 cohorts of 70 YBC pts candidates for BRCA gene testing treated at COEI in Mexico between 2013 to 2015 and IOCN from Romania between 2015 to 2016.Patient demographic and clinico-pathological information was recorded, and included: age at diagnosis, laterality presentation, clinical stage, Scarff-Bloom-Ricardson (SBR) grading, hormone receptor (HR), HER2neu and Ki 67% status Results: The analysis of these cohorts revealed clinical stage to be statistically significant different,(p = 0.0084), the Rom...

Journal of Clinical Oncology, 2017

e13116 Background: To determine types and frequencies of BRCA 1 (B1) or BRCA2 (B2) mutations in h... more e13116 Background: To determine types and frequencies of BRCA 1 (B1) or BRCA2 (B2) mutations in high-risk Romanian breast cancer patients, as there is no data published in this population. Methods: This prospective study evaluates the germline BRCA1/BRCA2 mutations in 200 Romanian high-risk breast cancer patients tested between February 2015-January 2017 at IOCN. Inclusion criteria selected patients diagnosed with breast cancer before 40 years, triple negative breast cancer under the age of 50, or having conventional family history criteria. All patients signed an informed consent. BRCA1/BRCA2 testing was performed using an AmpliSeq-based sequencing analysis, on the Ion Torrent Personal Genome Machine (Life Technologies) at RCFG. The pathogenic mutations were validated using Sanger technology. MLPA was performed for all 200 patients. Results: We analyzed 200 high-risk breast cancer patients and found 32 (16%) patients with pathogenic mutations, 23 (11.5%) patients with B1 and 9 (4.5...

Image and Vision Computing, 2018

In video surveillance, pedestrian attributes are defined as semantic descriptors like gender, clo... more In video surveillance, pedestrian attributes are defined as semantic descriptors like gender, clothing or accessories. In this paper, we propose a CNN-based pedestrian attribute-assisted person re-identification framework. First we perform the attribute learning by a part-specific CNN to model attribute patterns related to different body parts and fuse them with low-level robust Local Maximal Occurrence (LOMO) features to address the problem of the large variation of visual appearance and location of attributes due to different body poses and camera views. Our experiments on three public benchmarks show that the proposed method improves the state of the art on attribute recognition. Then we merge the learned attribute CNN embedding with another identification CNN embedding in a triplet structure to perform the person re-identification task. Both CNNs are pre-trained in a supervised way on attributes and person identities respectively, and then continue the training with a combined architecture for re-identification. We experimentally show that this fusion of "identity and attributes features" improves the overall re-identification.

Electrochimica Acta, 2018

A new synthesized molecular receptor, 2,2'-(ethane-1,2-diylbis((2-(azulen-2-ylamino)-2-oxoethyl)a... more A new synthesized molecular receptor, 2,2'-(ethane-1,2-diylbis((2-(azulen-2-ylamino)-2-oxoethyl)azanediyl))diacetic acid (L) has been used as building block for the development of new colorimetric and electrochemical sensors for metal ions complexation. The monomer L shows a high selectivity towards mercury ions among other cations, as highlighted by naked eye and UV-Vis spectra. The oxidative electropolymerization of L on glassy carbon electrodes lead to complexing polyL coated electrodes. They have been characterized by electrochemical techniques, their surface morphology and topography being evaluated by scanning electron microscopy and atomic force microscopy. These new

Materials and Corrosion, 2018

The aim of this study is to investigate the corrosion behavior of anodized Zr in oral cavity envi... more The aim of this study is to investigate the corrosion behavior of anodized Zr in oral cavity environment simulated by Afnor saliva as a function of the amount of dissolved oxygen. For comparison, non treated Zr was also investigated in the same conditions. In order to study the corrosion behavior at open circuit potential versus time, cyclic potentiodynamic polarization (CPP) and electrochemical impedance spectroscopy (EIS) were performed. scanning electron microscopy (SEM) with energy dispersive X-ray (EDX) imaging was used to observe the morphology and topography of the modified electrodes. Experimental data show that the decrease of oxygen concentration leads to the inhibition of the corrosion process for non treated Zr and Zr covered with nanotubes. A range of electrochemical stability of zirconia nanotubes in artificial Afnor saliva as a function of oxygen content in simulated oral cavity environment was established.

Sensors and Actuators B: Chemical, 2017

Abstract Metal ion complexing modified electrodes were obtained by oxidative electropolymerizatio... more Abstract Metal ion complexing modified electrodes were obtained by oxidative electropolymerization of 1 H -pyrrole-1-hexanoic acid (monomer L ) on glassy carbon disks with very good reproducibility in the elaboration process. Poly L films were characterized by scanning electron microscopy, atomic force microscopy, electrochemical techniques, and their hydrophilic character was evaluated by contact angle measurements. Using chemical preconcentration-anodic stripping technique at optimized parameters (pH, open circuit preconcentration time, reduction potential and time) linear calibration curves with ranges from 3 × 10 −11 M to 8 × 10 −8 M, and from 5 × 10 −11 M to 7 × 10 −8 M were obtained for Pb(II) ions using two thickness of poly L films. Their corresponding detection limits (S/N = 3) were 20pM and 30pM, respectively. These limits are favorable in comparison with other modified electrodes developed for Pb(II) detection. Poly L modified electrodes were used for electroanalysis of Pb(II) ion in tap waters, and the results were validated by GF-AAS method. Also, the modified electrodes presented very good stability over time.

Lecture Notes in Computer Science, 2018

Person re-identification consists in matching images of a particular person captured in a network... more Person re-identification consists in matching images of a particular person captured in a network of cameras with non-overlapping fields of view. The challenges in this task arise from the large variations of human appearance. In particular, the same person could show very different appearances from different points of view. To address this challenge, in this paper we propose an Orientation-Specific Convolutional Neural Network (OSCNN) framework which jointly performs body orientation regression and extracts orientation-specific deep representations for person re-identification. A robust joint embedding is obtained by combining feature representations under different body orientations. We experimentally show on two public benchmarks that taking into account body orientations improves the person re-identification performance. Moreover, our approach outperforms most of the previous state-of-the-art reidentification methods on these benchmarks.

arXiv (Cornell University), Dec 6, 2019

In most machine learning algorithms, training data is assumed to be independent and identically d... more In most machine learning algorithms, training data is assumed to be independent and identically distributed (iid). When it is not the case, the algorithm's performances are challenged, leading to the famous phenomenon of catastrophic forgetting. Algorithms dealing with it are gathered in the "Continual Learning" research field. In this paper, we study the regularization based approaches to continual learning and show that those approaches can not learn to discriminate classes from different tasks in an elemental continual benchmark: the class-incremental scenario. We make theoretical reasoning to prove this shortcoming and illustrate it with examples and experiments. Moreover, we show that it can have some important consequences on continual multi-tasks reinforcement learning or in pre-trained models used for continual learning. We believe that highlighting and understanding the shortcomings of regularization strategies will help us to use them more efficiently.

Pharmaceutics

Non-steroidal anti-inflammatory drugs (NSAIDs) showed effects in some hyperproliferative dermatol... more Non-steroidal anti-inflammatory drugs (NSAIDs) showed effects in some hyperproliferative dermatologic pathologies. The aim of the study is the assessment of anti-psoriasis effect of diclofenac and celecoxib using a mice tail model. The topical application of substances on the proximal mice tails was performed for two weeks. The effects on the epidermal granular layer and mean epidermal thickness (excluding the stratum corneum) were evaluated using hematoxylin–eosin staining. Orthokeratosis degree and percentual drug activity were calculated. A positive control group treated with tretinoin and two negative controls (white soft paraffin and untreated mice) were used. Orthokeratosis degree significantly increased in all the NSAIDs groups (celecoxib 1%, 2% and diclofenac 1%, 2%) and in the tretinoin 0.05% group, versus negative controls. Celecoxib 1% and 2%, tretinoin 0.05% and white soft paraffin significantly increased mean epidermal thickness, versus untreated mice. The values obtain...

Within the last two decades, the world has seen many financial crises which have spread around th... more Within the last two decades, the world has seen many financial crises which have spread around the globe and affected both developing and developed countries. Although each had its own particularities, there are some similar patters that stand out, most being related to an over evaluated asset in a country or region and a sharp withdrawal of funds that led to the burst of the crisis. Looking into each of them, there are some common causes and addressing these may help prevent any future financial shocks.

2018 25th IEEE International Conference on Image Processing (ICIP), 2018

Person re-identification is an important task in video surveillance systems. It consists in match... more Person re-identification is an important task in video surveillance systems. It consists in matching an image of a probe person among a gallery image set of people detected from a network of surveillance cameras with non-overlapping fields of view. The main challenge of person re-identification is to find image representations that are discriminating the persons' identities and that are robust to the viewpoint, body pose, illumination changes and partial occlusions. In this paper, we proposed a metric learning approach based on a deep neural network using a novel loss function which we call the Rank-Triplet loss. This proposed loss function is based on the predicted and ground truth ranking of a list of instances instead of pairs or triplets and takes into account the improvement of evaluation measures during training. Through our experiments on two person re-identification datasets, we show that the new loss outperforms other common loss functions and that our approach achieves state-of-the-art results on these two datasets.

Coatings

High-entropy alloys (HEAs), also known as multicomponent or multi-principal element alloys (MPEAs... more High-entropy alloys (HEAs), also known as multicomponent or multi-principal element alloys (MPEAs), differ from traditional alloys, which are usually based only on one principal element, in that they are usually fabricated from five or more elements in large percentages related to each other, in the range of 5%–35%. Despite the usually outstanding characteristics of HEAs, based on a properly selected design, many such alloys are coated with advanced composites after their elaboration to further improve their qualities. In this study, 73Ti-20Zr-5Ta-2Ag samples were covered with chitosan and a mixture of chitosan, bioglass, and ZnO particles to improve the materials’ antibacterial properties. A variety of methods, including scanning electron microscopy, atomic force microscopy, and mechanical and electrochemical determinations, has permitted a quantified comparison between the coated and uncoated surfaces of this medium–high-entropy alloy. The materials’ properties were enhanced by th...

Lecture Notes in Computer Science, 2019

We present a new replay-based method of continual classification learning that we term "condition... more We present a new replay-based method of continual classification learning that we term "conditional replay" which generates samples and labels together by sampling from a distribution conditioned on the class. We compare conditional replay to another replay-based continual learning paradigm (which we term "marginal replay") that generates samples independently of their class and assigns labels in a separate step. The main improvement in conditional replay is that labels for generated samples need not be inferred, which reduces the margin for error in complex continual classification learning tasks. We demonstrate the effectiveness of this approach using novel and standard benchmarks constructed from MNIST and FashionMNIST data, and compare to the regularization-based elastic weight consolidation (EWC) method [17, 34].

The electrochemical stability of CoCr based alloys is due to the passive oxide film grown on thei... more The electrochemical stability of CoCr based alloys is due to the passive oxide film grown on their surface, but it is to notice the aspect of ion release in their working time in this paper a new CoCrMo dental alloy alloy with small amount of niobium and zirconium (6% Nb and 0.8% Zr respectively) is modified by incorporation Ag on its surface. This paper aims to compare the ion release from CoCrMoNbZr in Tani Zuchi artificial saliva at various periods of time before and after Ag nanoparticles incorporation by pulselectrodeposition. The surface analysis reveled that Ag are nanoparticles with 100 nm average diameter. The ion release determination with inductively plasma mass spectrometer (ICP-MS established that after Ag incorporation the surface is more protective and Zr release decreased significantly. Electrochemical tests in artificial saliva Tani Zuchi sustain the ICP-MS determination which indicate better performance of alloy CoCr MoNbZr after Ag incorporation.

Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, 2018

In video surveillance, pedestrian attributes such as gender, clothing or hair types are useful cu... more In video surveillance, pedestrian attributes such as gender, clothing or hair types are useful cues to identify people. The main challenge in pedestrian attribute recognition is the large variation of visual appearance and location of attributes due to different poses and camera views. In this paper, we propose a neural network combining high-level learnt Convolutional Neural Network (CNN) features and low-level handcrafted features to address the problem of highly varying viewpoints. We first extract low-level robust Local Maximal Occurrence (LOMO) features and learn a body part-specific CNN to model attribute patterns related to different body parts. For small datasets which have few data, we propose a new learning strategy, where the CNN is pre-trained in a triplet structure on a person re-identification task and then fine-tuned on attribute recognition. Finally, we fuse the two feature representations to recognise pedestrian attributes. Our approach achieves state-of-the-art results on three public pedestrian attribute datasets.

Plastic and Reconstructive Surgery - Global Open, 2018

surgical technical difficulties, less complications and long term rates of survival.

Remote Sensing, 2021

Automatic ship detection provides an essential function towards maritime domain awareness for sec... more Automatic ship detection provides an essential function towards maritime domain awareness for security or economic monitoring purposes. This work presents an approach for training a deep learning ship detector in Sentinel-2 multi-spectral images with few labeled examples. We design a network architecture for detecting ships with a backbone that can be pre-trained separately. By using self supervised learning, an emerging unsupervised training procedure, we learn good features on Sentinel-2 images, without requiring labeling, to initialize our network’s backbone. The full network is then fine-tuned to learn to detect ships in challenging settings. We evaluated this approach versus pre-training on ImageNet and versus a classical image processing pipeline. We examined the impact of variations in the self-supervised learning step and we show that in the few-shot learning setting self-supervised pre-training achieves better results than ImageNet pre-training. When enough training data ar...

2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, 2021

We introduce in this paper a novel active learning algorithm for satellite image change detection... more We introduce in this paper a novel active learning algorithm for satellite image change detection. The proposed solution is interactive and based on a question & answer model, which asks an oracle (annotator) the most informative questions about the relevance of sampled satellite image pairs, and according to the oracle's responses, updates a decision function iteratively. We investigate a novel framework which models the probability that samples are relevant; this probability is obtained by minimizing an objective function capturing representativity, diversity and ambiguity. Only data with a high probability according to these criteria are selected and displayed to the oracle for further annotation. Extensive experiments on the task of satellite image change detection after natural hazards (namely tornadoes) show the relevance of the proposed method against the related work.

Materials and Corrosion, 2018

The paper studies the electrochemical stability in physiological serum and cell response of nanos... more The paper studies the electrochemical stability in physiological serum and cell response of nanostrucures obtained on Zr by anodizing in electrolytes containing fluorinated ethylene glycol and various amounts of K 2 CO 3. SEM analysis revealed that the porous nanostructure formation is greatly influenced by the various K 2 CO 3 concentrations in the electrolyte by tipping the balance between oxide formation and dissolution during the anodizing process. The porosities of the samples were evaluated from polarization resistances recorded in physiological serum. Chronoamperometric measurements were employed to establish a range of compactness for the samples. It was found that the amount of K 2 CO 3 from the electrolyte is directly responsible in obtaining oxide strata more resistant to corrosion. All tests performed on the samples such as corrosion, hydrophylicity and cell viability with MC3T3-E1 pre-osteoblasts cells from mouse bones revealed that the best coating for future research regarding nanostructured Zr implants was the one obtained from the electrolyte containing the highest amount of K 2 CO 3 .

Information Fusion, 2020

Continual learning (CL) is a particular machine learning paradigm where the data distribution and... more Continual learning (CL) is a particular machine learning paradigm where the data distribution and learning objective change through time, or where all the training data and objective criteria are never available at once. The evolution of the learning process is modeled by a sequence of learning experiences where the goal is to be able to learn new skills all along the sequence without forgetting what has been previously learned. CL can be seen as an online learning where knowledge fusion needs to take place in order to learn from streams of data presented sequentially in time. Continual learning also aims at the same time at optimizing the memory, the computation power and the speed during the learning process. An important challenge for machine learning is not necessarily finding solutions that work in the real world but rather finding stable algorithms that can learn in real world. Hence, the ideal approach would be tackling the real world in a embodied platform: an autonomous agent. Continual learning would then be effective in an autonomous agent or robot, which would learn autonomously through time about the external world, and incrementally develop a set of complex skills and knowledge. Robotic agents have to learn to adapt and interact with their environment using a continuous stream of observations. Some recent approaches aim at tackling continual learning for robotics, but most recent papers on continual learning only experiment approaches in simulation or with static datasets. Unfortunately, the evaluation of those algorithms does not provide insights on whether their solutions may help continual learning in the context of robotics. This paper aims at reviewing the existing state of the art of continual learning, summarizing existing benchmarks and metrics, and proposing a framework for presenting and evaluating both robotics and non robotics approaches in a way that makes transfer between both fields easier. We put light on continual learning in the context of robotics to create connections between fields and normalize approaches.

Journal of Clinical Oncology, 2018

e13572Background: Breast cancer (BC) is the most common malignancy in women with 6.6% of cases di... more e13572Background: Breast cancer (BC) is the most common malignancy in women with 6.6% of cases diagnosed in young women below the age of 40. Although there are significant similarities in demographic and clinico-pathological characteristic between geographical areas, the striking difference is in peak age for young populations. Objetive:Comparison of two young breast cancer(YBC) patients (pts) cohorts candidates for BRCA gene testing of two Developing Countries Methods: In this retrospective study, we analyzed 2 cohorts of 70 YBC pts candidates for BRCA gene testing treated at COEI in Mexico between 2013 to 2015 and IOCN from Romania between 2015 to 2016.Patient demographic and clinico-pathological information was recorded, and included: age at diagnosis, laterality presentation, clinical stage, Scarff-Bloom-Ricardson (SBR) grading, hormone receptor (HR), HER2neu and Ki 67% status Results: The analysis of these cohorts revealed clinical stage to be statistically significant different,(p = 0.0084), the Rom...

Journal of Clinical Oncology, 2017

e13116 Background: To determine types and frequencies of BRCA 1 (B1) or BRCA2 (B2) mutations in h... more e13116 Background: To determine types and frequencies of BRCA 1 (B1) or BRCA2 (B2) mutations in high-risk Romanian breast cancer patients, as there is no data published in this population. Methods: This prospective study evaluates the germline BRCA1/BRCA2 mutations in 200 Romanian high-risk breast cancer patients tested between February 2015-January 2017 at IOCN. Inclusion criteria selected patients diagnosed with breast cancer before 40 years, triple negative breast cancer under the age of 50, or having conventional family history criteria. All patients signed an informed consent. BRCA1/BRCA2 testing was performed using an AmpliSeq-based sequencing analysis, on the Ion Torrent Personal Genome Machine (Life Technologies) at RCFG. The pathogenic mutations were validated using Sanger technology. MLPA was performed for all 200 patients. Results: We analyzed 200 high-risk breast cancer patients and found 32 (16%) patients with pathogenic mutations, 23 (11.5%) patients with B1 and 9 (4.5...

Image and Vision Computing, 2018

In video surveillance, pedestrian attributes are defined as semantic descriptors like gender, clo... more In video surveillance, pedestrian attributes are defined as semantic descriptors like gender, clothing or accessories. In this paper, we propose a CNN-based pedestrian attribute-assisted person re-identification framework. First we perform the attribute learning by a part-specific CNN to model attribute patterns related to different body parts and fuse them with low-level robust Local Maximal Occurrence (LOMO) features to address the problem of the large variation of visual appearance and location of attributes due to different body poses and camera views. Our experiments on three public benchmarks show that the proposed method improves the state of the art on attribute recognition. Then we merge the learned attribute CNN embedding with another identification CNN embedding in a triplet structure to perform the person re-identification task. Both CNNs are pre-trained in a supervised way on attributes and person identities respectively, and then continue the training with a combined architecture for re-identification. We experimentally show that this fusion of "identity and attributes features" improves the overall re-identification.

Electrochimica Acta, 2018

A new synthesized molecular receptor, 2,2'-(ethane-1,2-diylbis((2-(azulen-2-ylamino)-2-oxoethyl)a... more A new synthesized molecular receptor, 2,2'-(ethane-1,2-diylbis((2-(azulen-2-ylamino)-2-oxoethyl)azanediyl))diacetic acid (L) has been used as building block for the development of new colorimetric and electrochemical sensors for metal ions complexation. The monomer L shows a high selectivity towards mercury ions among other cations, as highlighted by naked eye and UV-Vis spectra. The oxidative electropolymerization of L on glassy carbon electrodes lead to complexing polyL coated electrodes. They have been characterized by electrochemical techniques, their surface morphology and topography being evaluated by scanning electron microscopy and atomic force microscopy. These new

Materials and Corrosion, 2018

The aim of this study is to investigate the corrosion behavior of anodized Zr in oral cavity envi... more The aim of this study is to investigate the corrosion behavior of anodized Zr in oral cavity environment simulated by Afnor saliva as a function of the amount of dissolved oxygen. For comparison, non treated Zr was also investigated in the same conditions. In order to study the corrosion behavior at open circuit potential versus time, cyclic potentiodynamic polarization (CPP) and electrochemical impedance spectroscopy (EIS) were performed. scanning electron microscopy (SEM) with energy dispersive X-ray (EDX) imaging was used to observe the morphology and topography of the modified electrodes. Experimental data show that the decrease of oxygen concentration leads to the inhibition of the corrosion process for non treated Zr and Zr covered with nanotubes. A range of electrochemical stability of zirconia nanotubes in artificial Afnor saliva as a function of oxygen content in simulated oral cavity environment was established.

Sensors and Actuators B: Chemical, 2017

Abstract Metal ion complexing modified electrodes were obtained by oxidative electropolymerizatio... more Abstract Metal ion complexing modified electrodes were obtained by oxidative electropolymerization of 1 H -pyrrole-1-hexanoic acid (monomer L ) on glassy carbon disks with very good reproducibility in the elaboration process. Poly L films were characterized by scanning electron microscopy, atomic force microscopy, electrochemical techniques, and their hydrophilic character was evaluated by contact angle measurements. Using chemical preconcentration-anodic stripping technique at optimized parameters (pH, open circuit preconcentration time, reduction potential and time) linear calibration curves with ranges from 3 × 10 −11 M to 8 × 10 −8 M, and from 5 × 10 −11 M to 7 × 10 −8 M were obtained for Pb(II) ions using two thickness of poly L films. Their corresponding detection limits (S/N = 3) were 20pM and 30pM, respectively. These limits are favorable in comparison with other modified electrodes developed for Pb(II) detection. Poly L modified electrodes were used for electroanalysis of Pb(II) ion in tap waters, and the results were validated by GF-AAS method. Also, the modified electrodes presented very good stability over time.

Lecture Notes in Computer Science, 2018

Person re-identification consists in matching images of a particular person captured in a network... more Person re-identification consists in matching images of a particular person captured in a network of cameras with non-overlapping fields of view. The challenges in this task arise from the large variations of human appearance. In particular, the same person could show very different appearances from different points of view. To address this challenge, in this paper we propose an Orientation-Specific Convolutional Neural Network (OSCNN) framework which jointly performs body orientation regression and extracts orientation-specific deep representations for person re-identification. A robust joint embedding is obtained by combining feature representations under different body orientations. We experimentally show on two public benchmarks that taking into account body orientations improves the person re-identification performance. Moreover, our approach outperforms most of the previous state-of-the-art reidentification methods on these benchmarks.

arXiv (Cornell University), Dec 6, 2019

In most machine learning algorithms, training data is assumed to be independent and identically d... more In most machine learning algorithms, training data is assumed to be independent and identically distributed (iid). When it is not the case, the algorithm's performances are challenged, leading to the famous phenomenon of catastrophic forgetting. Algorithms dealing with it are gathered in the "Continual Learning" research field. In this paper, we study the regularization based approaches to continual learning and show that those approaches can not learn to discriminate classes from different tasks in an elemental continual benchmark: the class-incremental scenario. We make theoretical reasoning to prove this shortcoming and illustrate it with examples and experiments. Moreover, we show that it can have some important consequences on continual multi-tasks reinforcement learning or in pre-trained models used for continual learning. We believe that highlighting and understanding the shortcomings of regularization strategies will help us to use them more efficiently.

Pharmaceutics

Non-steroidal anti-inflammatory drugs (NSAIDs) showed effects in some hyperproliferative dermatol... more Non-steroidal anti-inflammatory drugs (NSAIDs) showed effects in some hyperproliferative dermatologic pathologies. The aim of the study is the assessment of anti-psoriasis effect of diclofenac and celecoxib using a mice tail model. The topical application of substances on the proximal mice tails was performed for two weeks. The effects on the epidermal granular layer and mean epidermal thickness (excluding the stratum corneum) were evaluated using hematoxylin–eosin staining. Orthokeratosis degree and percentual drug activity were calculated. A positive control group treated with tretinoin and two negative controls (white soft paraffin and untreated mice) were used. Orthokeratosis degree significantly increased in all the NSAIDs groups (celecoxib 1%, 2% and diclofenac 1%, 2%) and in the tretinoin 0.05% group, versus negative controls. Celecoxib 1% and 2%, tretinoin 0.05% and white soft paraffin significantly increased mean epidermal thickness, versus untreated mice. The values obtain...

Within the last two decades, the world has seen many financial crises which have spread around th... more Within the last two decades, the world has seen many financial crises which have spread around the globe and affected both developing and developed countries. Although each had its own particularities, there are some similar patters that stand out, most being related to an over evaluated asset in a country or region and a sharp withdrawal of funds that led to the burst of the crisis. Looking into each of them, there are some common causes and addressing these may help prevent any future financial shocks.

2018 25th IEEE International Conference on Image Processing (ICIP), 2018

Person re-identification is an important task in video surveillance systems. It consists in match... more Person re-identification is an important task in video surveillance systems. It consists in matching an image of a probe person among a gallery image set of people detected from a network of surveillance cameras with non-overlapping fields of view. The main challenge of person re-identification is to find image representations that are discriminating the persons' identities and that are robust to the viewpoint, body pose, illumination changes and partial occlusions. In this paper, we proposed a metric learning approach based on a deep neural network using a novel loss function which we call the Rank-Triplet loss. This proposed loss function is based on the predicted and ground truth ranking of a list of instances instead of pairs or triplets and takes into account the improvement of evaluation measures during training. Through our experiments on two person re-identification datasets, we show that the new loss outperforms other common loss functions and that our approach achieves state-of-the-art results on these two datasets.