3lLobo's repositories
ComputerVisionLab1
Photometric Stereo and Color in Computer Vision
ComputerVisionLab3
Edge detection
Advanced-Line-Finding-Project
This project detects road lanes using image preprocessing and computer vision object detection. Keeping track of the angle and curvature of the lane let's us predict the behaviour of the lanes and is matched with the result of step one to reduce the impact of erroneous lane detection.
BasicProbabilityTheroyUvA
Homework
ExtendedKalmanFilter-ObjectTracking
Implementation of the Extendet Kalman Filter on lidar and radar traffic data.
Git_puzzle
Puzzle quiz for isobar
MachineLearningLab2
UVA ML1 lab2
MachineLearningLab3
Lab 3 at UVA covering Kernel Methods
ModelPredictiveControl-SCD
Project on Model Predictive Control of Self-Driving Cars.
Natural_Language_Processing
Repo for work related with NLP
PID-Controller_SDC
Implementation of a PID controller for simulated cars. Tested on a simulated environment of a circular race track. Acceleration and steering is controlled by the PID controller.
QualitativeReasoning
University project Artificial Inteligence
ReinforcementLearning
Collection of my work on Reinforcement Learning, starting with the UvA course RL.
ReinforcementLearningAndPlanning
Advanced RL course applied to Planning and Optimization
SARSAvsQLearning
Python implementation of Q-learning and SARSA on a snake game
SDC-Path-Planning
C++ code to drive the autonomous car in the Udacity simulator. Features: avoid collision, adjust speed, change lane (left&right)
semantic-segmentation
This project labels the pixels of a road in images using a Fully Convolutional Network (FCN).
Unscented-Kalman-FIlter
The UKF uses multiple datapoints to approximate the Gaussian distribution of the next timestep. The points are weighted, thus unscented.
uvadlc_practicals_2019
Lab assignments for the Deep Learning course at the University of Amsterdam, April-May 2019
Vehicle-Detection-and-Tracking
Project to detect and track vehicles on video. A classifier detects vehicles and draws bounding boxes. HOG features are used to predict the position in the next frame.