Annotate datasets with a semi-trained or fully trained YOLOv5 model

Last update: May 14, 2022

Overview

YOLOv5 Auto Annotator

Annotate datasets with a semi-trained or fully trained YOLOv5 model

Prerequisites

Ubuntu >=20.04
Python >=3.7

System dependencies

sudo apt install python3-dev python3-pip

Python dependencies

cycler==0.11.0
fonttools==4.29.0
kiwisolver==1.3.2
lxml==4.6.4
numpy==1.21.4
opencv-contrib-python==4.5.5.62
opencv-python==4.5.5.62
packaging==21.3
Pillow==9.0.0
pyparsing==3.0.7
python-dateutil==2.8.2
six==1.16.0
tqdm==4.62.3

Install with the following command -

pip3 install -r requirements.txt

Run the application

Execute annotate.py in the following format -

usage: annotate.py [-h] [--viewmode] [--imgdir IMGDIR] [--annodir ANNODIR] [--confThreshold CONFTHRESHOLD] [--nmsThreshold NMSTHRESHOLD] [--width WIDTH] [--height HEIGHT] [--onnx_path ONNX_PATH] [--labels_path LABELS_PATH]

optional arguments:
  -h, --help            show this help message and exit
  --viewmode            Toggle View Mode
  --imgdir IMGDIR       Directory of images
  --annodir ANNODIR     Directory of annotations
  --confThreshold CONFTHRESHOLD
                        Class confidence
  --nmsThreshold NMSTHRESHOLD
                        NMS threshold
  --width WIDTH         Width of network input
  --height HEIGHT       Height of network input
  --onnx_path ONNX_PATH
                        Path to onnx file
  --labels_path LABELS_PATH
                        Path to labels file

Example -

python3 annotate.py --imgdir /home/kn1ght/Documents/images --annodir annotations --onnx_path models/YOLOv5s/yolov5s.onnx --labels_path models/YOLOv5s/coco.names --viewmode

Annotate datasets with a semi-trained or fully trained YOLOv5 model

Related tags

Overview

YOLOv5 Auto Annotator

Prerequisites

System dependencies

Python dependencies

Run the application

Owner

Akash James

(ICONIP 2020) MobileHand: Real-time 3D Hand Shape and Pose Estimation from Color Image

Codebase for the paper titled "Continual learning with local module selection"

A motion tracking system for any arbitaray points in a video frame.

Global Pooling, More than Meets the Eye: Position Information is Encoded Channel-Wise in CNNs, ICCV 2021

PyTorch Code for the paper "VSE++: Improving Visual-Semantic Embeddings with Hard Negatives"

Paper: De-rendering Stylized Texts

This is the code for "HyperNeRF: A Higher-Dimensional Representation for Topologically Varying Neural Radiance Fields".

🛰️ Awesome Satellite Imagery Datasets

The lightweight PyTorch wrapper for high-performance AI research. Scale your models, not the boilerplate.

The code for paper Efficiently Solve the Max-cut Problem via a Quantum Qubit Rotation Algorithm

Replication attempt for the Protein Folding Model

Anderson Acceleration for Deep Learning

4th place solution to datafactory challenge by Intermarché.

Computer-Vision-Paper-Reviews - Computer Vision Paper Reviews with Key Summary along Papers & Codes

RATCHET is a Medical Transformer for Chest X-ray Diagnosis and Reporting

Punctuation Restoration using Transformer Models for High-and Low-Resource Languages

The code for MM2021 paper "Multi-Level Counterfactual Contrast for Visual Commonsense Reasoning"

NLMpy - A Python package to create neutral landscape models

Learning to Prompt for Vision-Language Models.

FastyAPI is a Stack boilerplate optimised for heavy loads.