CDIoU and CDIoU loss is like a convenient plug-in that can be used in multiple models. CDIoU and CDIoU loss have different excellent performances in several models such as Faster R-CNN, YOLOv4, RetinaNet and . There is a maximum AP improvement of 1.9% and an average AP of 0.8% improvement on MS COCO dataset, compared to traditional evaluation-feedback modules. Here we just use as an example to illustrate the code.

Overview

CDIoU-CDIoUloss

CDIoU and CDIoU loss is like a convenient plug-in that can be used in multiple models. CDIoU and CDIoU loss have different excellent performances in several models such as Faster R-CNN, YOLOv4, RetinaNet and . There is a maximum AP improvement of 1.9% and an average AP of 0.8% improvement on MS COCO dataset, compared to traditional evaluation-feedback modules. Here we just use as an example to illustrate the code.

Control Distance IoU and Control Distance IoU Loss Function

by Chen Dong, Miao Duoqian

Introduction

Numerous improvements for feedback mechanisms have contributed to the great progress in object detection. In* this paper, we first present an evaluation-feedback module, which is proposed to consist of evaluation system and feedback mechanism. Then we analyze and summarize the disadvantages and improvements of traditional evaluation-feedback module. Finally, we focus on both the evaluation system and the feedback mechanism, and propose Control Distance IoU and Control Distance IoU loss function (or CDIoU and CDIoU loss for short) without increasing parameters or FLOPs in models, which show different significant enhancements on several classical and emerging models. Some experiments and comparative tests show that coordinated evaluation-feedback module can effectively improve model performance. CDIoU and CDIoU loss have different excellent performances in several models such as Faster R-CNN, YOLOv4, RetinaNet and ATSS. There is a maximum AP improvement of 1.9% and an average AP of 0.8% improvement on MS COCO dataset, compared to traditional evaluation-feedback modules.

There are some potential defects in the current mainstream target detection

  • It relies too much on the deepening of the backbone to extract features, so as to improve the accuracy of target detection;

  • The deepening of neural network, especially the deepening of backbone and neck, results in huge parameters and flops of the model;

  • Compared with the evaluation system (IoUs, the common ones are IoU and GIoU). At present, some new model optimization focuses more on the feedback mechanism (IoU losses), such as IoU loss, smooth loss, GIoU loss,CIoU loss, DIoU loss.

We propose Control Distance IoU and Control Distance IoU Loss Function (CDIoU and CDIoU loss for short).

Analysis of traditional IoUs and loss functions

  • Analysis of traditional IoUs

  • IoU: Smooth L1 Loss and IoU Loss

  • GIoU and GIoU Loss

  • DIoU loss and CIoU Loss

For more information, see Control Distance IoU and Control Distance IoU Loss Function for Better Bounding Box Regression

Installation

CDIoU and CDIoU loss is like a convenient plug-in that can be used in multiple models. CDIoU and CDIoU loss have different excellent performances in several models such as Faster R-CNN, YOLOv4, RetinaNet and ATSS. There is a maximum AP improvement of 1.9% and an average AP of 0.8% improvement on MS COCO dataset, compared to traditional evaluation-feedback modules. Here we just use ATSS as an example to illustrate the code.

These models use different frameworks, and some even have versions, so no code is provided in this article.


This ATSS implementation is based on FCOS and maskrcnn-benchmark and the installation is the same as them. Please check INSTALL.md for installation instructions.

ATSS bridges the gap between anchor-based and anchor-free detection via adaptive training sample selection. Comparison tests on ATSS exclude the essential interference between anchor-based and anchor-free detection. In these tests, the interference of positive and negative sample generation is eliminated, which give tests based on ATSS more representativeness.

CDIoU and CDIoU loss functions

CDIoU

Turning

For more information, see Control Distance IoU and Control Distance IoU Loss Function for Better Bounding Box Regression

Experiments

In order to verify the effectiveness of CDIoU and CDIoU loss in object detection, experiments are designed and applied to numerous models in this paper. These models encompass existing classical models and emerging models, reflecting certain robustness and wide adaptability.

image-20210315210941666

image-20210315214303929

For more information, see Control Distance IoU and Control Distance IoU Loss Function for Better Bounding Box Regression

Models

For your convenience, we provide the following trained models. All models are trained with 16 images in a mini-batch and frozen batch normalization (i.e., consistent with models in FCOS and maskrcnn_benchmark).

Model Multi-scale evaluation system feedback mechanism AP (val) AP (test-dev) pth
ATSS R 50 FPN 1x + CDIoU & loss NO CDIoU CDIoU loss 39.5 39.4 ATSS R 50 FPN 1x + CDIoU & loss
ATSS dcnv2 R 50 FPN 1x + CDIoU & loss NO CDIoU CDIoU loss 43.1 43.1 ATSS dcnv2 R 50 FPN 1x + CDIoU & loss
ATSS dcnv2 R 101 FPN 2x + CDIoU & loss NO CDIoU CDIoU loss 46.3 46.4 ATSS dcnv2 R 101 FPN 2x + CDIoU & loss
ATSS X 101 32x8d FPN 2x + CDIoU & loss NO CDIoU CDIoU loss 45.1 45.2 ATSS X 101 32x8d FPN 2x + CDIoU & loss
ATSS dcnv2 X 101 32x8d FPN 2x + CDIoU & loss NO CDIoU CDIoU loss 48.1 47.9 ATSS dcnv2 X 101 32x8d FPN 2x + CDIoU & loss
ATSS dcnv2 X 101 32x8d FPN 2x(MS) + CDIoU & loss YES CDIoU CDIoU loss 50.9 50.7 ATSS dcnv2 X 101 32x8d FPN 2x(MS) + CDIoU & loss

[1] The testing time is taken from FCOS, because our method only redefines positive and negative training samples without incurring any additional overhead. [2] 1x and 2x mean the model is trained for 90K and 180K iterations, respectively. [3] All results are obtained with a single model and without any test time data augmentation such as multi-scale, flipping and etc.. [4] dcnv2 denotes deformable convolutional networks v2. Note that for ResNet based models, we apply deformable convolutions from stage c3 to c5 in backbones. For ResNeXt based models, only stage c4 and c5 use deformable convolutions. All models use deformable convolutions in the last layer of detector towers. [5] The model ATSS_dcnv2_X_101_64x4d_FPN_2x with multi-scale testing achieves 50.7% in AP on COCO test-dev. Please use TEST.BBOX_AUG.ENABLED True to enable multi-scale testing.

MSCOCO test-dev

image-20210315222453294

Tips to improve performances

  • Floating learning rate

It is a consensus that the learning rate decreases as the iterative process in the experiment. Further, this paper proposes to check the loss every K iterations and increase the learning rate slightly, if the loss function does not decrease continuously. In this way, the learning rate will decrease and float appropriately at regular intervals to promote the decrease of the loss function.

  • Automatic GT clustering analysis

It is well known that AP can be effectively improved by performing cluster analysis on GT in the original dataset. We adjust anchor sizes and aspect ratios parameters based on the results of this cluster analysis. However, we do not know the number of clusters through the current approach. The main solution is to keep trying the number of clusters N , and then judge by the final result AP. Obviously, this exhaustive method takes a lot of time.

Contributing to the project

Any pull requests or issues are welcome.

CItations

Please cite our paper in your publications if it helps your research: And is not true!!

<_>This reference stuff is for fun only!!!!!!!!!

@inproceedings{chen2021CDIoU,
  title     =  {Control Distance IoU and Control Distance IoU Loss Function for Better Bounding Box Regression},
  author    =  {Chendong, Miaoduoqian.},
  booktitle =  {ICCV},
  year      =  {2021}
}
Owner
Alan D Chen
UJN : bachelor : MATH &CS | TONGJI : PhD : CV
Alan D Chen
Discord bot that shows valorant your daily store by using the Ingame API

Valorant store checker - Discord Bot Discord bot that shows valorant your daily store by using the Ingame API. written using Python and the Pycord lib

STACIA 226 Jan 02, 2023
Python script for download course from platzi.com

Platzi Downloader Tool Esta es una pequeña herramienta que hace mucho y que te ahorra una gran cantidad de trabajo a la hora de descargar cursos de Pl

Devil64-Dev 21 Sep 22, 2022
Lamblayer: a minimal deployment tool for AWS Lambda layers

lamblayer lamblayer is a minimal deployment tool for AWS Lambda layers. lamblayer does, Create a Layers of built pip-installable python packages. Crea

Yusuke Takahashi 2 Aug 19, 2022
A file-based quote bot written in Python

Let's Write a Python Quote Bot! This repository will get you started with building a quote bot in Python. It's meant to be used along with the Learnin

1 Nov 01, 2021
A Python interface between Earth Engine and xarray for processing weather and climate data

wxee What is wxee? wxee was built to make processing gridded, mesoscale time series weather and climate data quick and easy by integrating the data ca

Aaron Zuspan 160 Dec 31, 2022
A Slash Commands Discord Bot created using Pycord!

Hey, I am Slash Bot. A Bot which works with Slash Commands! Prerequisites Python 3+ Check out. the requirements.txt and install all the pakages. Insta

Saumya Patel 18 Nov 15, 2022
An API wrapper around Discord API written in Python

Diskord This library is a maintained fork of now archived library, discord.py. A modern and easy to use API wrapper around Discord API written in Pyth

Diskord 36 Aug 22, 2022
Send embeds using your discord personal account

Welcome to Embed Sender 👋 Send embeds using your discord personal account Install pip install -r requirements.txt Usage Put your discord token in ./

SkydenFly 11 Sep 07, 2022
Pysauce is a Discord bot which utilizes the SauceNAO API to locate the source of images.

Pysauce Pysauce is a Discord bot which utilizes the SauceNAO API to locate the source of images. Use Pysauce has one public instance always running, i

Akira 2 Oct 04, 2022
Reddit bot that uses sentiment analysis

Reddit Bot Project 2: Neural Network Boogaloo Reddit bot that uses sentiment analysis from NLTK.VADER WIP_WIP_WIP_WIP_WIP_WIP Link to test subreddit:

TpK 1 Oct 24, 2021
REPO USERBOT YANG DIBUAT DARI BERBAGAI REPO USERBOT GITHUB.

Lord Userbot Userbot Yang Digunakan Untuk Bersenang-Senang Di Telegram Repo Lord Userbot Repo Yang Dibuat Alvin Dari Berbagai Repo Userbot Github CARA

Alvin 70 Jan 02, 2023
The Dolby.io Developer Days Getting Started with Media APIs Workshop repo.

Dolby.io Developer Days Media APIs Getting Started Application About this Workshop and Application This example is designed to get participants workin

Dolby.io Samples 2 Nov 03, 2022
Transcript-Extractor-Bot - Yet another Telegram Voice Recognition bot but using vosk and supports 20+ languages

transcript extractor Yet another Telegram Voice Recognition bot but using vosk a

6 Oct 21, 2022
Template to create a telegram bot in python

Template for Telegram Bot Template to create a telegram bot in python. How to Run First add src to PYTHONPATH: export PYTHONPATH=${PWD} Then run: pyt

Ali Hejazizo 12 Dec 24, 2022
You can submit any PR and have SWAGS. Happy Hacktoberfest !

Excluded project Repository 🔴 🔴 🔴 - PR limit is reached. Please use another Repository Hacktoberfest 2021 🎉 🗣 Hacktoberfest encourages participat

Hansajith 63 Oct 21, 2022
Hasan Can Kaya - Konusanlar Ticket Notifier

Hasan Can Kaya - Konusanlar Ticket Notifier This script sends a notification to any telegram chat/group/channel when added a new available ticket to b

omer citak 3 Jan 31, 2022
Bot that embeds a random hysterical meme from Reddit into your text channel as an embedded message, using an API call.

Discord_Meme_Bot 🤣 Bot that embeds a random hysterical meme from Reddit into your text channel as an embedded message, using an API call. Add the bot

2 Jan 16, 2022
Hacktoberfest2021 - Submit Just 4 PRs to earn SWAGS and Tshirts🔥

dont contribute in this repo, contribute only in below mentioned repo Special Note For Everyone ''' always make more then 4 pull request lets you have

Keshav Singh 820 Jan 02, 2023
Indian Space Research Organisation API With Python

ISRO Indian Space Research Organisation API Installation pip install ISRO Usage import isro isro.spacecrafts() # returns spacecrafts data isro.lau

Fayas Noushad 5 Aug 11, 2022
Draw your telegram bot in draw.io and generate its code

README Contents: Draw your bot Install requirements Registering a telegram bot Draw bot Start point Message block Single choice block Functions block

DENIS TSITKO 23 Nov 09, 2022