<h3>Project&nbsp;<a href="https://ods.ai/projects/lacmus">Lacmus</a></h3>

Every person going outdoors has a chance to get lost. Usually people get lost in the forest and wilderness. About 120 000 Persons are lost In Russia every year. More than 100 000 Persons are lost in the USA every year.

Usually a person, who is lost in wilderness, meets a number of risk factors, such as: Dehydration, Hypothermia, Trauma, Panic. Unfortunately, this person has a high chance of dying.

Various search and rescue teams around the world are looking for missing people. For searches, they often use unmanned aerial vehicles with which they perform aerial photography.

But... Analyzing photos also needs human resources. How can we speed up this process? We can use Neural Networks!

Our project, Lacmus, helps Search &amp; Rescue Squads to find people, who get lost, by analyzing photos from copters with Neural Networks. And now we are ready for The First Competition!

At the competition, we offer you to solve the problem of recognizing and detecting lost people in aerial images. It is a classical one class object detection task. We need to detect people in photos from copters with an estimated altitude about 50m (164 ft).

<img alt="" src="https://storage.yandexcloud.net/datasouls-ods/images/2021/07/01/346cf7dc-6335-4046-8564-d058081a5efb/data-example.jpeg" />

<h3>Task Definition</h3>

During the competition, you are invited to solve the&nbsp;one class object detection&nbsp;task.

<h3>Metrics</h3>

We considering two metrics:

<ul>
	<li><a href="https://jonathan-hui.medium.com/map-mean-average-precision-for-object-detection-45c121a31173" rel="nofollow">mAP @ IoU 50</a></li>
	<li><a href="https://en.wikipedia.org/wiki/F-score" rel="nofollow">F1 @ IoU 50</a></li>
</ul>

The implementation of these metrics can be found in the file&nbsp;<a href="https://github.com/lacmus-foundation/events/blob/master/cv-competition-1/metrics/metrics.py"><code>metrics.py</code></a>

<h3>Submit Format</h3>

The submission has&nbsp;<code>csv</code>&nbsp;format the following fields:&nbsp;<code>id,xmin,ymin,xmax,ymax,score</code>. Where:

<ul>
	<li>id - id of image (e.g.&nbsp;<code>1.jpg</code>&nbsp;has id 1,&nbsp;<code>2.jpg</code>&nbsp;has id 2 etc);</li>
	<li>xmin, ymin, xmax, ymax - bounding box coordinates;</li>
	<li>score - prediction confidence (float: [0..1])</li>
</ul>

The submission file must include a header.

Example:

<pre lang="csv">
<code class="language-python">id,xmin,ymin,xmax,ymax,score
0,127,2877,271,2921,0.589
0,1751,2111,1888,2222,0.111
1,583,1111,666,1211,0.799
...</code></pre>

ODS slack channel&nbsp;<a href="https://opendatascience.slack.com/archives/CJ6UPQ3FZ">#ml4sg</a>

Every person has a chance to get lost. Usually people get lost in the forest or wilderness. At the competition, we offer you to solve the problem of recognizing and detecting lost people in aerial images.

Public

Baseline

Lacmus Foundation

Lacmus is a cross-platform software, which realize human detection on photos from unmanned aerial vehicle (UAV).

We are made automation of task to search the humans on set of photos from UAV.

We analyze photos from UAV using Neural Networks, Deep Learning &amp; Computer Vision to find people, who are lost in the wild forest.

This technology helps Search &amp; Rescue Services to find lost people in short time.

Increasing chances to resque a lost person.

Lacmus app should be installed on drone operator&#39;s notebooks to process a set of photos, downloaded from UAV after flight above people Search &amp; Rescue region.

Also we are working about building pipeline, which will help us to enlarge our dataset by using a number of photos from different Search &amp; Rescue drone operators from different regions.

Habr article [RU]:

&nbsp;<a href="https://habr.com/ru/company/ods/blog/483616/">https://habr.com/ru/company/ods/blog/483616/</a>

Datafest-2019 speech [RU]:

&nbsp;<a href="https://youtu.be/RQrjbAUoPNk?t=6270">https://youtu.be/RQrjbAUoPNk?t=6270</a>

Github:

&nbsp;<a href="https://github.com/lacmus-foundation/lacmus">https://github.com/lacmus-foundation/lacmus</a>

Telegram &nbsp;(Easy Way to Join Us) :

<a href="https://t.me/found_lacmus">https://t.me/found_lacmus</a>

We are waiting for ML researchers, backend/frontend developers and others with expertize in Computer Vision &amp; Deep Learning

&nbsp;

Members of our foundation:

gosha20777, kseniia, ei-grad, balezz, sergey_vakhreev, stasy, appetito, wanteich, Rumata, Palladdiumm, Silkindm, sergsh, sharov_am, nvsit, zvasil

e.t.c

<figure class="image"><img alt="https://raw.githubusercontent.com/lacmus-foundation/lacmus/master/docs/imgs/examples/04.png" src="https://raw.githubusercontent.com/lacmus-foundation/lacmus/master/docs/imgs/examples/04.png" />
<figcaption>BBoxed people on photo from UAV</figcaption>
</figure>

Andrew Grigorev

Palladdiumm

Sergey Gordeev

Jilly Ush

Sharov Anton

AndrejLabintsev

Yuriy Homyakov

Kseniia Terekhova

Gorgy Perevozchikov

The open-source project engaged in the search & rescue for missing peo

ML for Social Good

Lacmus

User Agreement

I accept <a target="_blank" rel="noopener noreferrer" href="https://drive.google.com/file/d/1n-fPczc0u7vYGvOzUUGH55HCbuRlCp8x/view?usp=sharing">Terms of User Agreement</a>

Privacy Policy

I accept <a target="_blank" rel="noopener noreferrer" href="https://drive.google.com/file/d/1JifpdktW5xCMAYgxzomLhYQZuxSrp0Wi/view?usp=sharing">Privacy Policy</a>

Consent

I give my <a target="_blank" rel="noopener noreferrer" href="https://storage.yandexcloud.net/datasouls-ods/DataFest2021_pres/Consent.pdf">consent</a> to «Sorevnovaniya Analiza Dannykh» LLC to process and transfer my personal data (name, email, address) to the partners of the event for information interaction

#	Teams (12)	MAP Score	F1 Score	Last	Entries
1	Sergey Zlobin	0.975934059	0.8335832084	3y	16
2	Slava Dodatko	0.967593858	0.9778565102	3y	5
3	Vadym Kozlovskyi	0.966741276	0.9778565102	3y	1
4	Maxim Lyuzin	0.9614938751	0.9139515455	3y	28
5	Yauhen Ustsinovich	0.9576977076	0.9181434599	3y	19
6	Vasiliy Adeshkin	0.9175074573	0.5641558442	3y	15
7	Alina Shchukina	0.8444422829	0.8578901482	3y	11
8	Maria Potapova	0.0001027071	0.0053835801	3y	1
9	Egor Pustovit	0.0001027071	0.0053835801	3y	1
10	Alexey Voytsekhovskiy	0.0001027071	0.0053835801	3y	1
11	Dmitrii Kutsenko	0.0001027071	0.0053835801	3y	1
12	Мария Филиппова	0.0001027071	0.0053835801	3y	4