My Portfolio

A showcase of My Projects and my abilities

Speaker Detection System
Using Lips


Speaker Detection is a image processing project in which our system will indentify whether a person in a video is speaking or not using lips landmarking. In this system facial landmarking using Dlib is used to detect lips and then widh and height is calculated. If width and height is changing person is speaking otherwise not speaking. this is the module of a Speaker Diarization using image processing project which i have been working on.

Logo Recognition
system

Logo Recognition is the project where we needed aur system to recognize Tv channel through its logo. This project is built using OpenCv,Python and algorithm used in this project is Structural Similarity Index Measurement(SSIM).




Speaker Diarization (Speaker Identification)

In this project we used audio dataset from joe rogan podcast in which edward snowden is appeared as guest we then convert audio to Spectrogram using librosa and save images of audio after 5 seconds then build two classes one for joe rogan and another for edward then put different spectrogram from different audio to detect edward snowden and joe rogan

Voilence Detection


In this Project we used google images to make our dataset we used images of accidents, buildings on fire, cars on fire etc to build our dataset then we used deeplearning to train our model to detect voilence




QR Generator and Recognizer

QR Generator and Recognizer is small application built using PYQT5 python pyzbar to generate Qr by taking input from user and recognize them afterward according to requirement.



Boats Recognition
System

In this project drone fly over the dock and take snaps of the boats docked there and then over system will detect boats recognize boats Using QR code or OCR which is on the front of the boat.System will Identify unknown and known boats and the number of boats docked there.The boats recognition system is built using python, OpenCv Imutils, Pyzbar.

Audio Video Auto Transcription and analysis Software

In this software user will input audio or video and software will transcribe its audio into text form for further analysis. User can enter keywords to find in the audio if those keywords are present in the video or audio the Software will return that line with timestamp in which those keywords appeared and if user press the play button user can also hear that specific part of the audio. User can also save the project so that he/she don't have to transcribe it again

Scheduler (Desktop App)


Scheduler Desktop app is built using python. In this software Admin select people who need vacation. Admin then put date and name into software and then software check the maxout limit for the people on and between that days and if approved software automatically send email to Admin, Employee and manager.