KEMAL KILIÇASLAN

Artificial Intelligence Developer

I am Kemal, I live in Ankara, Turkey. I graduated from Kastamonu University, Department of Mathematics in 2022 with a GPA of 2.8/4. Currently, I am focusing on improving my technical skills in areas such as artificial intelligence, machine learning, deep learning, and image processing by applying the knowledge from my mathematics background. To achieve this, I am exploring AI algorithms through personal projects, online courses, and practical applications specific to image processing in line with my interest. These projects include tasks such as image detection, recognition, and interpretation. Through this hands-on work, I gain experience that helps me deepen my understanding of AI concepts and further develop my technical skills. My goal is to continue developing these skills and contribute to AI solutions in real-world contexts.

PROJECTS

6174 Kaprekar Constant with Python Programming

6174 Python code

6174 app calculator

A 4 digit number with at least one digit different from the one selected by the user is selected, the number 6174 is reached in maximum 7 steps by applying consecutive arithmetic operations.

Face Detection and Person Recognition

Detect Faces in Photo

Detect Faces in Video

Real Time Face Detection

Persons Recognition in Photo

Persons Recognition in Video

Real Time Person Recognition

It is a face detection and person recognition project on photos, videos and snapshots using Haarcascade classifier algorithm.

Road Lane Lines Detection

Road Lane Lines

Detected Road Lane Lines

It is a project to detect lane lines on roads for autonomous vehicles in which artificial intelligence is actively involved.

Garbage Classification with Convolutional Neural Network (CNN)

Garbage Classification Class

Garbage Classification Predicted Class

It is a classification project by Stanford University students using Convolutional Neural Network (CNN) model on TrashNet dataset consisting of 6 different classes.

Object Detection and Segmentation with YOLOv8

Object Detection Image

Object Detection Video

Real Time Object Detection

Object Segmentation Image

Object Segmentation Video

Real Time Object Segmentation

Object detection & segmentation project from photographs, videos and snapshots with the 8th version of the YOLO algorithm.

Vehicle Recognition with Segmentation Training on a Custom Dataset

car

segmented car

pickup

segmented pickup

truck

segmented truck

It is a project to train the model with segmentation method on 20 randomly selected cars, pickups and trucks and to measure model success on 5 randomly selected cars, pickups and trucks.

Pose Detection with YOLOv8 using Wolfram Mathematica

Test Image

Heatmap Image

Skeleton Image

Skeleton Map Image

Microsoft COCO is a pose detection project that uses the latest version of YOLO (You Only Look Once) Version 8, YOLO models developed by Ultralytics in the dataset used for image recognition, segmentation, captioning, object detection and keypoint estimation consisting of more than three hundred thousand images.

Facial Expression Recognition

angry

disgust

fear

happy

neutral

sad

surprise

The FER-2013 dataset created for the Facial Expression Recognition Competition consists of 35887 images with a resolution of 48x48 pixels showing facial expressions corresponding to seven different emotion classes (Angry, Disgust, Fear, Happy, Sad, Surprise and Neutral). 80% of this dataset was used for training and 20% for testing.

Vehicle Speed Estimation

Vehicle Flow

ROI Area

Target Source

ROI Coordinates

Vehicle Speed Estimation

This project uses YOLO for object detection to track vehicles and estimate their speeds. Video frames are processed with thresholds and perspective adjustments via a ViewTransformer class. The output features tracking lines, bounding boxes, and velocity labels, enabling accurate speed estimation and object detection in defined regions.

Traffic Signs Recognition for Turkey

Traffic Signs in Turkey

Traffic Signs Recognition for Turkey

This project detects 17 common traffic signs in city centers for autonomous vehicles using a dataset of screenshots from Google Street View for Turkey and trains the sensitive detection and automation process with the YOLOv8x model on 668 images via Roboflow and creates an automation system for autonomous vehicles.

Data Visualization of Turkey Population with Plotly

Choropleth Map

Bar Graph

Stack Bar Graph

Donut & Pie Graph

It is a data visualization project on the map with line, bar, stack bar, pie, donut charts and choropleth method with male, female and total population data from 1927 to 2023 taken from TURKSTAT using Plotly and Folium libraries.