<feed xmlns="http://www.w3.org/2005/Atom"> <id>https://kcnote.github.io/</id><title>KCnote Dev Blog</title><subtitle>Personal dev blog about C++, computer vision, algorithms, and AI</subtitle> <updated>2026-06-10T14:18:02+08:00</updated> <author> <name>Steve Cho</name> <uri>https://kcnote.github.io/</uri> </author><link rel="self" type="application/atom+xml" href="https://kcnote.github.io/feed.xml"/><link rel="alternate" type="text/html" hreflang="en" href="https://kcnote.github.io/"/> <generator uri="https://jekyllrb.com/" version="4.4.1">Jekyll</generator> <rights> © 2026 Steve Cho </rights> <icon>/assets/img/favicons/favicon.ico</icon> <logo>/assets/img/favicons/favicon-96x96.png</logo> <entry><title>Image Segmentation</title><link href="https://kcnote.github.io/posts/AI-Pytorch-05-Image-Segmentation/" rel="alternate" type="text/html" title="Image Segmentation" /><published>2026-06-09T23:00:00+08:00</published> <updated>2026-06-09T23:00:00+08:00</updated> <id>https://kcnote.github.io/posts/AI-Pytorch-05-Image-Segmentation/</id> <content type="text/html" src="https://kcnote.github.io/posts/AI-Pytorch-05-Image-Segmentation/" /> <author> <name>kang</name> </author> <category term="Artificial Intelligence" /> <category term="Artificial Intelligence - PyTorch" /> <summary>Image Segmentation Prerequisites 1. Image Segmentation 1.1 Example Train Segmentation # ========================= # 1. Install # ========================= !pip install -q transformers datasets evaluate accelerate tqdm import torch import numpy as np import matplotlib.pyplot as plt from datasets import load_dataset from torch.utils.data import DataLoader from transformers import Segf...</summary> </entry> <entry><title>Object Detection YOLO</title><link href="https://kcnote.github.io/posts/AI-Pytorch-04-Object-Detection-YOLO/" rel="alternate" type="text/html" title="Object Detection YOLO" /><published>2026-06-07T23:00:00+08:00</published> <updated>2026-06-07T23:00:00+08:00</updated> <id>https://kcnote.github.io/posts/AI-Pytorch-04-Object-Detection-YOLO/</id> <content type="text/html" src="https://kcnote.github.io/posts/AI-Pytorch-04-Object-Detection-YOLO/" /> <author> <name>kang</name> </author> <category term="Artificial Intelligence" /> <category term="Artificial Intelligence - PyTorch" /> <summary>Object Detection YOLO Prerequisites 1. Object Detection YOLO 1.1 Example Convert Data for YOLO import shutil from pathlib import Path import pandas as pd from sklearn.model_selection import train_test_split from tqdm import tqdm ROOT_DIR = Path(__file__).resolve().parents[1] # ComputerVision RAW_DIR = ROOT_DIR / "data" / "gwhd_2021" IMAGE_DIR = RAW_DIR / "images" CSV_PATH = RAW_...</summary> </entry> <entry><title>Defect Inspection</title><link href="https://kcnote.github.io/posts/AI-Pytorch-03-Defect-Inspection/" rel="alternate" type="text/html" title="Defect Inspection" /><published>2026-06-04T23:00:00+08:00</published> <updated>2026-06-08T22:56:59+08:00</updated> <id>https://kcnote.github.io/posts/AI-Pytorch-03-Defect-Inspection/</id> <content type="text/html" src="https://kcnote.github.io/posts/AI-Pytorch-03-Defect-Inspection/" /> <author> <name>kang</name> </author> <category term="Artificial Intelligence" /> <category term="Artificial Intelligence - PyTorch" /> <summary>Defect Inspection Prerequisites 1. Defect Inspection 1.1 Example import os import glob import cv2 import torch import timm import numpy as np from PIL import Image from tqdm import tqdm from sklearn.metrics import roc_auc_score from torchvision import transforms DEVICE = "cuda" if torch.cuda.is_available() else "cpu" BASE_DIR = os.path.dirname(os.path.abspath(__file__)) DATA_ROOT =...</summary> </entry> <entry><title>Publisher and Subscriber</title><link href="https://kcnote.github.io/posts/ROS-01-Publisher-and-Subscriber/" rel="alternate" type="text/html" title="Publisher and Subscriber" /><published>2026-06-02T23:00:00+08:00</published> <updated>2026-06-02T23:00:00+08:00</updated> <id>https://kcnote.github.io/posts/ROS-01-Publisher-and-Subscriber/</id> <content type="text/html" src="https://kcnote.github.io/posts/ROS-01-Publisher-and-Subscriber/" /> <author> <name>kang</name> </author> <category term="Robotics" /> <category term="Robotics - ROS2" /> <summary>Publisher and Subscriber Prerequisites 1. ROS2 1.1 Install wsl --install -d Ubuntu-24.04 sudo apt update sudo apt upgrade -y 1.2 Workspace sudo apt install ros-jazzy-desktop -y echo “source /opt/ros/jazzy/setup.bash” » ~/.bashrc</summary> </entry> <entry><title>Convolution RGB with SIMD</title><link href="https://kcnote.github.io/posts/CODE-SIMD-02-Convolution-RGB-with-SIMD/" rel="alternate" type="text/html" title="Convolution RGB with SIMD" /><published>2026-05-25T23:00:00+08:00</published> <updated>2026-05-25T23:00:00+08:00</updated> <id>https://kcnote.github.io/posts/CODE-SIMD-02-Convolution-RGB-with-SIMD/</id> <content type="text/html" src="https://kcnote.github.io/posts/CODE-SIMD-02-Convolution-RGB-with-SIMD/" /> <author> <name>kang</name> </author> <category term="CODE" /> <category term="CODE - SIMD" /> <summary>Convolution RGB with SIMD Prerequisites SIMD 1. Naive Code About 5119 by 5119 image. Processing time: 305ms uint8_t* pU8SrcBuf = (uint8_t*)imgSrc.data; uint8_t* pU8DstBuf = (uint8_t*)imgDst.data; int32_t i32SteopRows = imgSrc.step; int32_t i32PixelBytes = imgSrc.elemSize(); for(int i32Row = 2; i32Row &amp;lt; imgSrc.rows - 2; ++i32Row) { uint8_t* pU8SrcBufRowR = pU8SrcBuf + i32Row *...</summary> </entry> </feed>
