Sara Saberi Moghadam Tehrani
Machine Learning Engineer
About Me
MSc. graduated in AI with 2+ years of experience in researching and building applied machine learning applications focused on web crawlers. Cooperation in smart search engine startup with 168,000 active users and 680,496 active products. I am interested in cutting-edge technology, researching, and applying AI algorithms to solve real-world problems.
Finally, I attempt to think critically and enjoy being creative, innovative, and team-oriented.
Bio
Professional Skills
Work Experience
SnapMode (Iranian Smart Fashion Search Engine SnapMode)
Designed and developed a Low Latency Scalable Focused Web Crawler to extract fashion data from E commerce websites using Apache Storm, Solr, Kafka and Milvus. (+10M Product Pages). Enhanced the content-based image retrieval Accuracy using a Triplet Generative Adversarial Networks (CBIR-GAN) to feature embedding.(82% accuracy on the in-shop products). Optimized the search performance of vector queries using clustered milvus.
Rasad (University News Analysis and Tracking System)
Designed and developed an news analysis and monitoring system that leveraged from BERT model for sentiment analysis and improved negative comments detection with 82% accuracy rate.(Focusing on university news)Portfolio
Apache Storm / Solr / Kafka / Milvus / Color Detection/ CNN Models / FastAPI/ Vuejs
Iranian Smart Fashion Search Engine
Designed and developed a Low Latency Scalable Focused Web Crawler to extract fashion data from E-commerce websites. Enhanced the content-based image retrieval Accuracy using a Triplet Generative Adversarial Networks (CBIR-GAN) to feature embedding.
Apache Storm / Solr / Sentiment Analysis / Twitter & Web Crawling / Bert NLP Model / FastAPI / Vuejs
Rasad (University News Analysis and Tracking System)
Designed and developed an news analysis and monitoring system that leveraged from BERT model for sentiment analysis and improved negative comments detection with 82% accuracy rate.(Focusing on university news)
GANs models / Wavelet Transform / Pytorch
Texture synthesis in image to image translation in the field of fashion AI
In this research, we had presented a generative model called WBT-GAN for texture synthesis problem, which was an extension of the existing Texture-GAN network using a four-level wavelet transform and error definition based on it in the objective function of the model.
Mask-RCNN Model / Wavelet Transform / Pytorch / Object Detection / Instance Segmentation
Improving Semantic Segmentation Perfomance in the Field of Fashion using Deep Learning
In this research We had improved mask localization accuracy in instance segmentation in COCO benchmark and Deep Fashion2 dataset. We introduced new architecture base on MaskR-CNN. We added a sub coeffeint wavelet improvment subnetwork to maintain more detail in produced mask.
Education
Publications
Submitted to Big Data Research Journal, October 2021.
11th International Conference on Computer and Knowledge Engineering (ICCKE), Mashhad, Iran, 2021, pp. 1-last page., October 2021.
Published in Journal of Medical Signals and Sensors, vol. 11, no. 3, pp. 159-168, July 2021.
in 42nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC’20), Montreal, Canada, pp. 1-last page. 2020.
In press
Conferences
Covid-19 Diagnosis / Image Classification / Deep CNN Models
Conference: 42nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC'20)At: Montreal, Canada
This paper analyses the ability of various deep learning strategies such as VGG19, Resnet101, VGG16, and InceptionResNetV2 in the classification of covid-infected patients from pneumonia and normal individuals using an exceptionally large dataset of 15,163 X-ray images, reporting 94% accuracy for the best model. Clinical Relevance—Results indicate the reliable capabilities of the deep learning techniques, such as the InceptionResnetV2 structure, for a robust identification and classification of covid-19-infected individuals using chest X-ray images.
GANs models / Wavelet Transform / Pytorch
11th International Conference on Computer and Knowledge Engineering (ICCKE 2021), October 28-29, 2021, Ferdowsi University of Mashhad
In this study, a generative model called WBT-GAN is proposed by using the four-level WT and employing objective function for defining of its loss function. In general, WBT-GAN is an extension of the existing network Texture-GAN. Experimental results showed that these changes have improved image resolution and sharpening and have led to better texture spread.
References
Dr. Reza Azmi
Associate Professor
Computer Engineering Faculty of Engineering University of Alzahra
Tehran, Iran
E-mail: azmi@alzahra.ac.ir
Narges Norouzi
Data Scientist, Toobatech Company
E-mail: na.norozi@gmail.com
Dr. Hamid Abbasi
Department of Engineering Science The University of Auckland
Auckland,New Zealand
E-mail: h.abbasi@auckland.ac.nz