Kiran Magdum

Detail-oriented team player with strong analytical and organizational skills. Ability to handle multiple projects simultaneously with a high degree of accuracy. Takes minimal time to Adapt to new technology and gives emphasis on learning, implementing and know-how sharing.

• Diabetic Retinopathy Detection Using CNN Architecture (Jan 2021 to June 2021)
  ● Designed an interface which takes retina fundus image as input and processes it in backend using 3 different CNN architectures and outputs the best possible result to identify the diabetic retinopathy stage. ●In this research, two deep CNN models were proposed with an ensemble technique to detect all the stages of DR by using balanced and imbalanced datasets. The models were trained with Kaggle dataset. Balanced dataset was wont to train both models, and that we test these models with balanced and imbalanced datasets.
  
• Mask and Social Distance Detection (Sep 2020 to Dec 2020)
  ● Development of a video analytical model for detecting masked faces and recognizing crowd gathering using Python module libraries: Keras, Tensorflow, Caffe, OpenCV.
  ● For Real Time Prediction and labeling of masked faces in frame, we proposed a solution on the use of MobileNetV2 Architecture for feature verification through Narrow->Wide->Narrow Approach.
  ● Implemented person class from COCO dataset as well as YOLO Framework for precise object detection.
  
• Fall Detection- by sensors and camera (Mar 2020 to July 2020 )
  ● Proposed system monitors an elderly person in real time and in case a fall is detected, an emergency number is contacted.
  ● Wearable device: We made use of Arduino UNO and a MPU 6050 sensor for fall recognition and packed in a single chip.
  ● For camera based fall detection: Implemented using USB camera and a raspberry pi 4 module comprising real time images of the elderly with a database consisting of images of persons in fallen or normal state.
  
• Design To develop MATLAB GUI/App for estimation of correlation and power spectral density of signals and verification of properties of correlation. (Aug 2019 to Nov 2019)
  ● Estimated concepts like correlation, PSD and its properties by using the MATLAB tool.
  ● The MATLAB algorithm is used to generate the cross spectral matrix by the Welch method with reasonable performance.
• Closed loop control of DC Motor (Mar 2019 to Jul 2020)
  ● The aim is to control the speed of the DC motor using the low cost data acquisition board i.e. the arduino board and simulation is performed in MATLAB Simulink support package.
  ● We used L298 Motor Driver, to maintain the desired speed on the generator operating on the same axis to the dc motor and to control the exceeding voltage and current.

Certificate of appreciation for technical paper presentation in IMCC's IC2TMA (international) conference on topic entitled as 'Face -Mask and Social Distance Detection System".

• Lead MAN India Engineering newsletter for all the four quarters of 2022
• Exhibition committee lead for MAN family day event.
• Volunteered as a teacher for computer literacy program arranged for senior citizens.
• Conducted blood donation camp in VIT, pune organized by BJGMC and SASSOON general hospital.

• English (C1)
• German (A1)
• Hindi
• Marathi
• Kannada