Research & Development of state-of-the art solutions with a focus to provide end-to-end solutions across the complete value chain encompassing original research, algorithm development, testing and production.
Combination of core research competency with multi-domain and diverse work environment exposure. Multi-modality Image Processing and Signal Processing, Medical Image Processing (CT, MRI, X-ray, Microscopic images), Computer Vision, Pattern Recognition, Machine Learning, Soft-computing, Computer Aided Diagnostic (CAD) Tools, Wavelets, Feature Extraction, Segmentation and Classification.
Advenio TecnoSys Pvt. LTd
From September 2010
Advenio is a multi-disciplinary Consultancy with a focus to provide end-to-end Solutions across the complete value chain encompassing original research, algorithm development, testing and production. Advenio functions with main aims: Innovation, Quality, Reliability and Cost-effectiveness to cater to Clients needs. Advenio consists of a compact dynamic, highly skilled and motivated group mastered in solving critical challenges. Team members are of multifarious disciplines and international exposure having academic backgrounds from top-notch institutions and robust Industry experience.
Advenio provides quality services and solutions to Image Processing, Machine Vision problems and applications with a specialized focus on Medical and Biomedical Image and Signal Processing, Industrial Imaging and Surveillance & Security. Notwithstanding that Advenio is skilled to provide solutions to any kind of imaging needs whatever be the application area.
• A wavelet based technique for noise reduction in Synthetic Aperture Radar (SAR) Images
• Automatic radar target classification using wavelets
• Image processing based automatic answer script evaluation from scanned images
• Prone/ Supine and Cranial/Caudal classification in full body CT images
• An unsupervised classification method to differentiate Sinus and Colon in full body CT images automatically for further processing in CAD applications
• A novel curvature feature extraction scheme to reduce false positives due to ribs and rib crossings in chest X-ray images which are further used for automatic identification of lung nodules
• Detection of aneurysms in non-contrast thick slice brain CT images
• Bird species identification by analyzing their song/chirp spectrograms
• Copy and move forgery detection in images: An image forensic application
June 2009 - August 2010
• Worked on several image processing algorithms like, Image Registration, Segmentation, and Feature Extraction on medical images of various modalities.
• Worked on detection and quantification of Tuberculosis Bacilli from sputum smears and chest X-rays.
• Worked on breast skin-line delineation and lesion detection in mammogram images.
• Worked on several business proposals for mammogram analysis in PACS environment.
Senior Member Technical Staff
Siemens Medical Solutions
May 2005 - May 2009
• Worked on detection of nodules in low resolution (large slice thickness) in Lung Computed Tomography images
• Designed several features to characterize pulmonary embolism (PE) and PE-look-alikes (looks similar to PE but are actually not) in CT pulmonary angiography (CTPA).
• Designed several feature extraction techniques for lesion (nodule) detection in Chest X-ray for improvement in an existing CAD prototype.
• Pioneered and designed image processing algorithms for automatic detection of different body parts in Computed Tomography (CT) images.
• Developed and designed an automatic technique for detection of intracranial hemorrhage and acute intracranial hemorrhage in various anatomical locations in brain Computed Tomography (CT) for trauma cases where no contrast can be applied and the CT has large slice thickness.
• Designed a technique for vessel tracing for locating proximal and distal vessels applied to Pulmonary Embolism (PE) severity scoring.
• Pioneered, motivated, supervised and managed invention disclosures amongst team members within the Business Unit of the organization as part of Innovation drive.
Scientist - D
Defense Research Development Organization
November 2002 - February 2004
• • Implemented designing, analysis and development of some new and novel algorithms for radar signal and image processing.
o Radar Target Identification and Classification
A generic algorithm is developed to identify radar targets. A high fidelity simulation of the signal samples produced by the radar targets has been created. The output of the simulation is designed for combination with the output of clutter and signal processor simulation so that target detection in clutter can be examined. The algorithm is based on wavelet transform. Characteristic features of the radar echo signal is extracted by using wavelet frames. The features so extracted are classified employing a feed-forward neural network.
o Enhancement of Synthetic Aperture Radar Image
A brief about some credible projects
Qualifications & Certifications
University of Pennsylvania
Indian Statistical Institute
University of Calcuatta
University of Calcutta
Skillpages has been acquired by Bark.com!
Bark.com is pioneering the way people find local services. Skillpages is the world’s premier directory of service providers.Find out more