The goal of this project is to create a specialized model for detecting and recognizing specific wild animals, including Elephant, Gorilla, Giraffe, Lion, Tiger, and Zebra. We gathered images of these animals and used the Roboflow annotation tool to manually label each animal class. After annotation, the data was exported in the YOLOv8 format.

Next, we trained a custom YOLOv8 model on this dataset to accurately detect and recognize the selected animal species in images. The project leverages YOLOv8’s object detection capabilities to improve detection accuracy for wildlife monitoring and research purposes.

You can find more details about the project on GitHub by clicking on this link. To view the training logs and metrics on wandb, click here.

A custom object detection dataset for vehicle and pedestrian tracking.
Includes annotated instances of mycar, person, car, and motorcycle.
This dataset is designed to train and evaluate real-time models (e.g. YOLOv8/YOLOv12) for tasks such as surveillance, traffic monitoring, or autonomous systems.

📦 License: MIT
📸 Source: Collected from private camera footage and public domain datasets
📊 Class Distribution: 4 classes across ~700 images
⚙️ Augmented via Roboflow with blur, scale, flip, and exposure variance

This project supports iterative model-assisted labeling using Roboflow Train and Deploy.
Optimized for model-assisted annotation — detect first, fix later!

The goal of this project is to build a specialized bird detection model capable of recognizing specific bird species, including Crow, Kingfisher, Myna, Owl, Peacock, Pigeon, and Sparrow. We compiled a diverse dataset of images containing these bird species and used the Roboflow annotation tool to manually label their locations within each image. Once the annotation was completed, the data was exported in the YOLOv8 format.

Next, we trained a custom YOLOv8 model on this dataset to accurately detect and identify the chosen bird classes across different environments and scenarios. This project harnesses the power of YOLOv8’s object detection capabilities to improve the precision and reliability of bird species identification in images.

You can find more details about the project on GitHub by clicking on this link.

🏏 Bowling Analysis & Ball Tracking Project 📌 Project Overview This project focuses on real-time ball and stump detection in cricket videos. The goal is to provide an AI-powered analysis of bowling performance by tracking the ball's speed, trajectory, and movement. The project leverages YOLOv8 for object detection and uses custom datasets to improve accuracy. The final application will be a Flutter-based mobile app, enabling seamless video analysis for bowlers at all levels.

🎯 Class Descriptions The dataset includes 5 key classes for training the YOLOv8 model: Ball - Tracks the cricket ball’s movement. Stumps - Detects all three stumps together. Stump - Detects individual stumps. Person - Identifies players in the video. Bat - Detects the presence of the cricket bat. These annotations are manually labeled using LabelImg and formatted for YOLOv8 training.

📅 Current Status & Timeline ✅ Project Setup & Planning (✅ Completed) ✅ Dataset Collection (✅ Completed) ✅ Data Annotation & Labeling (✅ Completed) ✅ YOLOv8 Model Training (🔄 Testing Different Configurations) 🚀 Final Model Optimization & Evaluation (🔜 Next Step) 📱 Flutter App Development & Integration (🔜 Final Phase)

🔗 External Resources 📺 YouTube Tutorial Followed: YOLOv8 Object Detection Guide 📄 Project Blog Updates: Bowling Analysis Blog 📂 GitHub Repository (To be added) 📖 YOLOv8 Documentation: Ultralytics YOLOv8 Docs 📝 Contribution & Labeling Guidelines

Dataset Standardization: Ensure class labels match across all datasets before training. Annotation Format: Use YOLO format [class_id x_center y_center width height] in .txt files. Merging Datasets: Keep consistent class IDs across multiple datasets to avoid conflicts. Training Best Practices: Train on diverse clips for better model generalization

Project Description: Celebrity Face Detection using YOLOv8

The goal of this project is to create a personalized face detection model for recognizing specific celebrities, including Virat Kohli, Maria Sharapova, Elon Musk, Katherine Langford, and Cristiano Ronaldo. We gathered images of these celebrities and used the Roboflow annotation tool to manually mark their faces. After annotation, the data was exported in the YOLOv8 format.

Next, we trained a custom YOLOv8 model on this dataset to identify the faces of the chosen celebrities. The project aims to accurately detect and recognize these celebrities' faces in images, utilizing the object detection capabilities of YOLOv8 to enhance face detection performance.

You can find the more details about the project on Github by clicking in this link. To view the training logs and metrics on wandb you can click here.

A custom object detection dataset for vehicle and pedestrian tracking.
Includes annotated instances of mycar, person, car, and motorcycle.
This dataset is designed to train and evaluate real-time models (e.g. YOLOv8/YOLOv12) for tasks such as surveillance, traffic monitoring, or autonomous systems.

📦 License: MIT
📸 Source: Collected from private camera footage and public domain datasets
📊 Class Distribution: 4 classes across ~700 images
⚙️ Augmented via Roboflow with blur, scale, flip, and exposure variance

This project supports iterative model-assisted labeling using Roboflow Train and Deploy.
Optimized for model-assisted annotation — detect first, fix later!

This project focuses on detecting and classifying plant growth stages using the YOLOv8 object detection model. The model is trained to recognize three distinct stages of plant development * Germination * Growing * Flowering

The dataset used in this project is a combination of publicly available images sourced from platforms like Roboflow Universe and Kaggle, as well as custom images captured manually to enhance dataset diversity and model accuracy. Proper credit is acknowledged for public image contributions.

Contains Images from Call of Duty Modern Warfare & Warzone gameplay and has labels for Enemy and Head.

Originally used to train a Yolov5 model to detect when enemies are in view and used a GIMX adapter with Python to send movement controls to connected PS4. Find the complete code on my Github.

This dataset can be used to train custom Computer Vision models to recognize when enemy players appear and locate them.

Checkout this video of the model running on a Twitch streamer's video (Faze Testy): https://youtu.be/cxFpTIK8aYE