PI-DECODER

Introduction

PI-DECODER, a decoder structure designed for Precise Iris Segmentation and Location. The decoder structure is shown below:

Please check technical paper.pdf in the "reference" subfolder for more details.

How to use?

For african dataset, you can enter the following script on your terminal:

python main.py --mode test --model_path ./models/african_best.pth --test_mode 1 --train_dataset african

Then you have iris mask, pupil mask and outer iris mask that are predicted by the input images. At the same time, the relevant index data will be displayed on your terminal.

(ijcb) PS F:\workspace\code\pytorch\PI-DECODER> python main.py --mode test --model_path ./models/african_best.pth --
test_mode 1 --train_dataset african
Namespace(batch_size=1, beta1=0.9, beta2=0.999, img_size=(640, 640), lr=0.0002, mode='test', model_path='./models/af
rican_best.pth', num_epochs=100, num_workers=2, result_path='./result/', test_mode=1, test_path='./dataset/test/', t
rain_dataset='african', train_path='./dataset/train/', valid_path='./dataset/valid/')
image count in train path :5
image count in valid path :5
image count in test path :40
Using Model: PI-DECODER
0.0688 seconds per image

----------------------------------------------------------------------------------------------------------------
|evaluation     |e1(%)          |e2(%)          |miou(%)        |f1(%)          |miou_back      |f1_back        |
----------------------------------------------------------------------------------------------------------------
|iris seg       |0.384026       |0.192013       |91.175200      |95.350625      |95.386805      |97.574698      |
|iris mask      |0.569627       |0.284813       |93.159855      |96.430411      |96.270919      |98.060105      |
|pupil mask     |0.078793       |0.039396       |93.138878      |96.409347      |96.529547      |98.184718      |
----------------------------------------------------------------------------------------------------------------
|average        |0.344149       |0.172074       |92.491311      |96.063461      |96.062424      |97.939840      |
----------------------------------------------------------------------------------------------------------------

Besides, if you don't have groud-truth files or just want to save the results, use test mode 2.

python main.py --mode test --model_path ./models/african_best.pth --test_mode 2 --train_dataset african

Requirements

The whole experiment was run on the NVIDIA RTX 3060. The following are recommended environment configurations.

matplotlib        3.3.4
numpy             1.19.5
opencv-python     4.5.1.48
pandas            1.1.5
Pillow            8.1.2
pip               21.0.1
pyparsing         2.4.7
python-dateutil   2.8.1
pytz              2021.1
scipy             1.5.4
setuptools        52.0.0.post20210125
six               1.15.0
thop              0.0.31.post2005241907
torch             1.7.0+cu110
torchstat         0.0.7
torchsummary      1.5.1
torchvision       0.8.1+cu110

precise iris segmentation

Related tags

Overview

PI-DECODER

Introduction

How to use?

Requirements

Owner

Universal Adversarial Triggers for Attacking and Analyzing NLP (EMNLP 2019)

Journey is a NLP-Powered Developer assistant

NLP-SentimentAnalysis - Coursera Course ( Duration : 5 weeks ) offered by DeepLearning.AI

This library is testing the ethics of language models by using natural adversarial texts.

Honor's thesis project analyzing whether the GPT-2 model can more effectively generate free-verse or structured poetry.

Guide to using pre-trained large language models of source code

The repository for the paper: Multilingual Translation via Grafting Pre-trained Language Models

A collection of Korean Text Datasets ready to use using Tensorflow-Datasets.

A modular framework for vision & language multimodal research from Facebook AI Research (FAIR)

VoiceFixer VoiceFixer is a framework for general speech restoration.

A crowdsourced dataset of dialogues grounded in social contexts involving utilization of commonsense.

Easy to use, state-of-the-art Neural Machine Translation for 100+ languages

LSTC: Boosting Atomic Action Detection with Long-Short-Term Context

ChessCoach is a neural network-based chess engine capable of natural-language commentary.

My Implementation for the paper EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Tasks using Tensorflow

Beyond Accuracy: Behavioral Testing of NLP models with CheckList

skweak: A software toolkit for weak supervision applied to NLP tasks

Unsupervised Language Model Pre-training for French

BERT score for text generation

Multilingual text (NLP) processing toolkit