| layout | page |
|---|---|
| title | Portfolio |
| permalink | /portfolio/ |
Focuses on identifying and segmenting road cracks using a UNet model. It provides various loss function implementations to evaluate performance. Learn more on GitHub.
Implements BiSeNet for real-time semantic segmentation of facial features, useful for applications like digital makeup and augmented reality. Learn more on GitHub.
This repository uses SCRFD for face detection and ArcFace for face recognition, supporting real-time inference from webcams and videos. Learn more on GitHub.
This repository contains code and instructions for performing object detection using YOLOv5 inference with ONNX Runtime. Learn more on GitHub.
This repository contains code and instructions for performing license plate detection using YOLOv5 inference with ONNX Runtime. Learn more on GitHub.
Reimplementation of DarkNet in PyTorch for flexibility and performance in deep learning. Learn more on GitHub.
FaceBoxes is a high-performance, real-time face detection model optimized for CPUs. This implementation provides efficient and accurate face detection without needing powerful GPUs. Learn more on GitHub.
This project implements the UNet architecture using PyTorch for image segmentation tasks. Trained on the Carvana dataset, it uses Dice loss and Cross-Entropy loss for training and offers high performance and flexibility. Learn more on GitHub.
This project implements fast neural style transfer using PyTorch. It applies artistic styles to images in real-time, leveraging the MSCOCO dataset for training and offering deployment options with Flask. Learn more on GitHub.
Reimplementation of "EAST: An Efficient and Accurate Scene Text Detector" using PyTorch. This project provides an efficient and accurate model for text detection in natural scenes, trained on the ICDAR2015 dataset. Learn more on GitHub.
A collection of tutorials and examples for learning PyTorch. It covers basic to advanced topics, helping users to understand and implement various deep learning models. Learn more on GitHub.
This project implements a captcha recognition system using PyTorch. It utilizes an RNN architecture for recognizing captchas and includes pre-trained weights, training scripts, and inference scripts. Learn more on GitHub.
This project generates images of Korean license plates with YOLO format labels using Python. It provides scripts for generating passenger car and truck license plate images, as well as tools for label distribution analysis. Learn more on GitHub.
This project provides tools to convert YOLO annotation format to Pascal VOC format and vice versa using Python. It includes scripts for both relative and absolute coordinate conversions, supporting flexible dataset management. Learn more on GitHub.
Implementation of YOLOv1 (Real-Time Object Detection) using PyTorch. This project includes training scripts, evaluation metrics, and pre-trained models, making it a comprehensive solution for object detection tasks. Learn more on GitHub.
Implementation of DeepLabV3 using PyTorch for semantic image segmentation. This project includes training scripts, evaluation metrics, and pre-trained models, providing a robust solution for segmentation tasks. Learn more on GitHub.
Implementation of deep text recognition using PyTorch. This project includes scripts for training, evaluation, and pre-trained models for recognizing text in images, leveraging advanced architectures like BiLSTM and attention mechanisms. Learn more on GitHub.
This project implements an OCR model for reading captchas using the Keras API. It combines CNN and RNN architectures, demonstrating a functional approach to solving captcha recognition tasks. Learn more on GitHub.
YOLOv1 re-implementation using PyTorch with ResNet50 as the backbone. This project provides scripts for training, evaluation, and detection, making it a robust solution for real-time object detection. Learn more on GitHub.
CRAFT (Character-Region Awareness for Text detection) implementation using PyTorch. This project includes pretrained models and scripts for text detection, providing robust performance in detecting text regions in images. Learn more on GitHub.
PyTorch implementation of MobileNetV3 for both large and small models. This project provides scripts for model training, evaluation, and inference, facilitating efficient deployment in various image classification tasks. Learn more on GitHub.
PyTorch implementation of MobileNetV2. This project includes training scripts, evaluation metrics, and pretrained models, providing efficient solutions for image classification tasks. Learn more on GitHub.