Undergrad Projects

Overview

An interactive game developed using opencv3 and python3. User can control the player by moving any blue object in front of camera. The camera will detect the object from background by image processing and calculate it's angle and movement and change the position of the player car accordingly.

Key Features

• • Use of OpenCV Library: Used OpenCV library for performing real-time blue color segmentation via HSV thresholding and contour detection to compute object centroid; updates directional movement flags based on position;
• • Full-fledged playable game: Developed a game loop using the PyGame library, featuring a car selection system and binary-based high score saving mechanism.

Repository

Description

Developed a YOLO(v1)-based detection model to classify the front and rear sides of various vehicles using a fully custom dataset. The dataset comprised 250 images captured via mobile phone, annotated using third-party software. An 80/20 train-test split was used. The project utilized Python 3, TensorFlow 1.0, NumPy, and OpenCV 3 as core dependencies.

Overview

Crime is the most chattered topic in recent days and its prevention is necessary. The essence of machine learning is to analyze data and extract useful information. It is a branch of artificial intelligence based on the idea that systems can learn from data, identify patterns and make decisions with minimal human intervention. Natural language processing, and deep learning algorithms used in this field also perform well. Data provides useful information to detect crime patterns. The aim of this paper is to build different crime prediction models, compare and analyze the results of the models. We applied machine learning techniques on Chicago crime dataset (2005-2017). Models predicts the primary crime type depending on some other features. Analyzing the results we visualize how accurately the machine learning models work. Model evaluation techniques are used to choose the best model and how well the models will perform in future. The soul purpose of this research work is to give the best idea of how machine learning can be used by the law enforcement agencies to detect, predict and solve crimes at a much faster rate.