Sumana Ramachandra

Detail-oriented, meticulous and dedicated software engineer with 4+ years of experience in software development for large-scale enterprise systems. A highly collaborative team player with excellent problem-solving abilities, who thrives in fast pace environments with a passion for delivering innovative solutions and transforming software solutions from conception to execution.

Block-Level Classification of Unstructured Data. (10/2020)

• Patent issuer and number US 121232.01 (Filed)

 "Double and triple compression-based forgery detection in JPEG images using deep convolutional, neural network”, Journal of Electronic Imaging March 2020 https://doi.org/10.1117/1.JEI.29.2.023006 

Deep learning based Melanoma Segmentation and Classification on Dermascopic Skin Lesion Images (02/2022 - Present)

Currently working on employing deep learning algorithms for segmenting specific region of interests such as lesions in dermascopic images. Involves image pre-processing,  exploring architecture modifications of ResUnet and Unet++ variants for medical image segmentation and extracting features and determining the severity of the disease by segmenting skin lesions and classifying them. 

Re-Compression based Image Forgery Detection Using Deep Neural Network (09/2018 - 05/2019)

Implemented a Convolutional Neural Network for detecting multiple compression and localizing the region of forgery done on JPEG images. The deployed CNN was trained with the extracted image compression artifacts and then tested against various sizes of forgery, implemented using MATLAB and Google TensorFlow in Python environment. This also includes detecting forgery in the presence of various post-processing attacks on the image under investigation with an average accuracy of 92%.

Online Library Management System (02/2015 - 05/2015)

Online Library Management System is a web-based application developed using PHP and MySql that deals with maintaining data about books and transactions such as issue, return, renewal of books, penalty charges and book replacement for Students and Professors managed by the admin. MySql is used as the database in this application. 

• Dell Release Contributions Award: Critical Contributions towards release (07/2021)
• Hackathon winner Best business value: File level in- flight IO encryption within block storage (11/2020)
• Dell Extraordinary Award: Contributions towards release (04/2020)
• Hackathon winner Cross platform mobility category: NextGen Multipathing Mechanism on SPDK (10/2019)