Virtual Ads или как прорекламировать Adidas в CS:GO

Всем привет, меня зовут Евгений Мунин. Я Senior ML Engineer в Ad Tech в платформе ставок для рекламы и автор ТГ канала ML Advertising . В данной статье мы поговорим об одном из способов повышения узнаваемости брендов в спорте, а точнее виртуальной рекламе. Разберем размещение рекламных баннеров на видео и напишем пример на Python и OpenCV, где разместим логотип Adidas с использованием алгоритма детектирования ключевых точек SIFT и гомографии для искажения баннера под перспективу.

https://habr.com/ru/articles/889386/

The OpenCV Perception Challenge for Bin-Picking is in full-swing, with teams submitting solutions to the leaderboard. If you're in the robotics or AI space, this is a great way to test your skills, win a share of the $60,000 in prizes, and most excitingly be part of an official CVPR Workshop at CVPR 2025 in Nashville!

Sign up to participate: https://bpc.opencv.org/web/challenges/challenge-page/1/overview
See the leaderboard: https://bpc.opencv.org/web/challenges/challenge-page/1/leaderboard/1

OpenCV is happy to announce we’ve once again been selected as a mentor organization for Google Summer of Code! Let's build something awesome together this summer https://summerofcode.withgoogle.com/programs/2025-ao/organizations/opencv #OpenCV #gsoc #GSoc2025

At 9am, in about 30 mins, we're live with the organizers of Chinatown Hacks, a hackathon being held in San Francisco's Chinatown this March. Learn about the event and the people behind it, live on YouTube https://www.youtube.com/watch?v=76R-mJvK524 #OpenCV #AI #SanFrancisco

The latest movement at the bird feeders. #CameraTrap #bluetit #greattit #nuthatch #Parakeet #Gatwick #RaspberryPi #opencv #picamera2 #python

We're live at 9am Pacific with Matt Flagg of the world's first autonomous motorsports team, Code 19! Come see the state of the art in robot race cars. https://www.youtube.com/live/Mc7c8WVGuCE #OpenCV #ComputerVision #Motorsports

How to segment X-Ray lungs using U-Net and Tensorflow

You can find link for the code in the blog : https://eranfeit.net/how-to-segment-x-ray-lungs-using-u-net-and-tensorflow/

Check out our tutorial here : https://youtu.be/-AejMcdeOOM&list=UULFTiWJJhaH6BviSWKLJUM9sg

Enjoy
Eran

roses are read,
violets are blue
submissions are open
so we're waiting for you

The testing phase is officially open for submissions in the OpenCV Perception Challenge for Bin-Picking (BPC), and more news on the blog. #OpenCV #ComputerVision #AI

https://opencv.org/blog/important-updates-bin-picking-challenge-phase-1-submissions-open-more/

This week we’re bringing you ANOTHER competition. The IEEE Low-Power Computer Vision Challenge is about improving the energy efficiency of computer vision for running on systems with stringent resource constraints. Watch live at 9am pacific to learn more https://youtube.com/live/8IgRNvBCjSc #OpenCV #ComputerVision

today in #fedora qa:
* looked into os-autoinst builds failures on f42 and #rawhide, they seem to be caused by #opencv 4.11.0 not working right on #s390x (bigendian) - poked into that a bit and filed a bug, https://github.com/opencv/opencv/issues/26913 , tried to do some bisecting but it's hard slogging so far
* restarted varnish on the staging proxy server (fixed access to various stg sites)
* helped fix the wiki styling problem - https://pagure.io/fedora-infrastructure/issue/12385
* merged and tested https://pagure.io/fedora-infra/ansible/pull-request/2462

I sat in 3 meetings with industrial vision vendors to discuss a project that requires both barcode, OCR and has to have a local monitor. All 3 vendors said it would require 2 cameras at $4k each + acc and no local monitor. "It just isn't possible on 1 camera!" I had a demo running on my laptop by the end of meeting with the 3rd vendor. So just for a demo. #raspberrypi 5, HQ camera, #opencv

Pokemon + Image Recognition fun at Michigan Python last night: https://www.youtube.com/watch?v=uGi9ZkiDVcw #Python #OpenCV #Pokemon

AI, FreeRTOS и Linux в кармане: возможности LicheeRV Nano

В этой статье речь пойдет о разработке под отладочную плату LicheeRV Nano - компактное устройство размером с две пятирублевые монеты, но обладающее впечатляющими возможностями. Плата способна одновременно запускать Linux и FreeRTOS , выполнять инференс нейронных сетей (будет разобран запуск YOLO в 100 FPS и LLama2.c) благодаря встоенному NPU с производительностью 1 TOPS , а также управлять периферийными устройствами: GPIO, I2C, UART, SPI, CSI камерой, Wi-Fi, Bluetooth и Ethernet. Это первая часть статьи, представляющая собой методическое руководство по работе с платой. Во второй части будет рассмотрена разработка полноценного проекта на её основе.

https://habr.com/ru/articles/880230/

#c++ #linux #электроника #программирование #npu #yolo #yolov8 #робототехника #llm #opencv

Registrations for the Perception Challenge for Bin-Picking have been open for about a month, and nearly 1000 participants have registered! Now it’s time for us to open submissions, and for you to show us what you can do. For today’s live stream we’ve invited Vahe Taamazyan from Intrinsic for a full tutorial on how to get the code, get the data, and submit your solutions. https://www.youtube.com/live/32D0TrsI06E

Ever wondered how to bring image recognition to the web? Come join us tomorrow (Thursday) at 7pm at Michigan Python!

I'll be sharing how to build a mobile image recognition app using OpenCV and PyScript using a fun Pokémon example. We'll cover:
- Building mobile-friendly computer vision apps
- Running Python in the browser with PyScript
- Real-world applications of OpenCV
- Live demo and code examples

#Python #OpenCV #PyScript #Pokemon

See you there!

https://www.meetup.com/michigan-python/events/305384906

So basically face recognision is: compare current #embedding distance to database of embeddings and closest is considered a match?
#facerecognition #opencv

TIL that a trivial Python3 OpenCV program (open an image, attempt to display it in a window) on a Raspberry Pi tries to open over 2800 shared libraries

(EDIT: more like nearly 1400, 'cos strace opens the symbolic link and the library ... but still)

strip-p-ed out

Some more pictures of the #patterns to bypass #cctv #facerecognition.

Picture 1: data visualisation of CCTVs ratios in 5 European cities (cctv/surface, cctv/population, cctv/crime index). #knitting on the #kniterate machine

Picture 2: doodled stripes on #inkscape with hatches paste effect. #knitting on the #kniterate machine

Picture 3: #sewing a ready-made pattern from deadstock fabrics

Picture 4: Exhibition view. A visitor trying the garment with live CCTV face recognition (#opencv)

Model: Élodie Goldberg
#fabricademy project 2024

For more info see my previous post!

We're live at 9am Pacific with a live, free, tutorial from OpenCV University. Learn how to make and use NeRFs, incredible hi-resolution, shareable representations of real-life items and spaces. https://youtube.com/live/Wvn3IzqVMJY #OpenCV #AI

We just released wasmVision 0.2 along with wasmCV 0.7 just in time for @fosdem

Come see our talk in the #webassembly dev room on Sunday at noon!

https://fosdem.org/2025/schedule/track/webassembly/