วีดีโอ

Thanks Dr. Satya! Thanks Phil! Thanks OpenCV community, we learned so much from OpenCV!

Please post link to Mini Pupper bot website. Thanks.

*Rerun: An Open-Source Tool for Visualizing Computer Vision and Robotics Data* * *0:27** Introduction to Rerun:* Rerun is an open-source SDK for logging and visualizing computer vision and robotics data, allowing developers to debug and understand their systems with minimal code. It supports various data types, including 3D point clouds, images, tensors, and scalars. * *0:33** Early Success:* Rerun gained significant traction within a week of its launch on GitHub, accumulating over 1,400 stars. * *7:46** Core Functionalities:* Rerun allows for powerful data inspection, including hovering over elements for detailed information, scrubbing through timelines, and creating interactive dashboards for complex pipelines. * *11:58** Availability and Licensing:* Rerun is available in beta, is open-source, and is offered under a dual license (MIT and Apache 2). Users can install it via pip (Python) or cargo (Rust). C and C++ support are planned for the future. * *14:39** Origins at Volumental:* Rerun's development stemmed from the founders' experience at Volumental, a 3D foot scanning company, where the need for effective visualization tools became apparent. * *26:31** Technical Overview:* The Rerun SDK is available in Python and Rust, with plans for C and C++ support. It utilizes a websocket connection to stream data to the viewer, enabling live visualization and analysis. Data can be saved to disk in .rrd (Rerun Data) files for later review and collaboration. * *28:16** Built with Rust:* Rerun leverages the performance, safety, and portability of the Rust programming language, allowing for a smooth 60 FPS visualization experience and compatibility with web assembly for browser-based usage. * *33:38** Gaming-Inspired Design:* Rerun's design philosophy draws inspiration from the gaming world, prioritizing performance, interactivity, and a fluid user experience. It incorporates concepts like immediate mode GUI and entity component systems for enhanced responsiveness and composability. * *36:16** Demo: Colmap 3D Reconstruction:* A demonstration showcasing the visualization of a 3D reconstruction generated using Colmap, highlighting Rerun's ability to handle point clouds, camera frustums, and key points. Users can customize the visualization by adjusting parameters such as point size and camera frustum distance. * *48:16** Demo: Stable Diffusion:* A demo visualizing the internal workings of a stable diffusion model, including the input image, depth map, text embeddings, latent representations, and final output. Rerun aids in understanding the complex processes within AI models. * *52:50** Demo: MRI Scan Visualization:* Rerun's ability to handle and visualize 3D volumetric data is demonstrated using an MRI scan. Users can slice through the volume in different directions and adjust the color mapping for better analysis. * *54:23** Demo: MediaPipe Pose Tracking:* Rerun visualizes the output of MediaPipe's pose tracking and segmentation, highlighting its effectiveness in analyzing real-time data streams. The example showed how Rerun can help identify issues like motion blur and missing data points. *Key Takeaways:* * Rerun is a powerful and versatile tool for visualizing and understanding complex data in computer vision and robotics applications. * Its open-source nature and intuitive interface make it accessible to a wide range of users. * The tool's performance and flexibility are enhanced by its Rust foundation and gaming-inspired design principles. * Rerun can be integrated into various workflows, including those involving popular libraries like PyTorch and TensorFlow. I used gemini-1.5-pro-exp-0801 to summarize the transcript. Cost (if I didn't use the free tier): $0.11 Input tokens: 30299 Output tokens: 803

Awesome session..

SM keeps surprising us with goodies! Thank you, Dr. Mallick and Phil.

all my greetings to you and all my respects in fact. I would like to ask you for a grievance. I'm having trouble downloading the opencv.js library that works with facial recognition.

all my greetings to you and all my respects in fact. I would like to ask you for a grievance. I'm having trouble downloading the opencv.js library that works with facial recognition.

all my greetings to you and all my respects in fact. I would like to ask you for a grievance. I'm having trouble downloading the opencv.js library that works with facial recognition.

all my greetings to you and all my respects in fact. I would like to ask you for a grievance. I'm having trouble downloading the opencv.js library that works with facial recognition.

all my greetings to you and all my respects in fact. I would like to ask you for a grievance. I'm having trouble downloading the opencv.js library that works with facial recognition.

Hhjjjgf

A Python plumber emitting comments about C++ 😂

What about classification, how would that be done? Let's say I have 3 categories: scratch, dirt, severe damage How can I classify that?

Ay how do I get on the next one? I'm ready to rock!

Very informative and entertaining. 🙂

Hello. This is an awesome development. Pls provide links to the resources in the description if possible. Thank you.

truly a visionary man

This is excellect. One thing will be possible in case of Industrial defect classification like pitting or corrosion it can tell us without supervised model right?

I need a job of annotator or data labelling in international company Do you help me?

Hi

India

Hi

Where to get Apo

What is log in - it is covered by avatar, thanks

why are there so many vehicle noises in the background

Awesome Work Katelyn Gan!! I'm very impressed with your project! You have a great future ahead of you! Keep it up! Also, thank you OpenCV for putting this together!

The zipline docuseries you showed in 1:05:31 is in Ghana. The drones deliver blood samples to remote areas as an emergency response for rural areas that cannot access health facilities.

😮😢😮😢😢😢😢😢😢😢😢😢😢😢😮😮😮😢😢😮😮😮😮😢😢

Also, thank you guys at OpenCV TH-cam Channel!

Thank you for putting this interview together! I enjoyed it so much! Murtaza Hassan is an incredible educator in Computer Vision, AI and Robotics!! I learned so much from him!

presentation starts at 12:25

😊

Badhiya

Jhukal

Congratulations for this awesome lecture! Just curious if you design the slides yourself, because they are amazing.

MAIN DESK TOP PE MEDIA RAKHHO DUNIYA KO DIKHVO 🔬 SCIENCE FREE POST 📪 NO CHARGES NO 🚫 COST DEAR SIR/ MEDAM GOOD AFTERNOON 🕒 HAPPY NEW YEARS 2024 ALL THE BEST 👍💯 BETTAR & BEAUTIFUL FOR ALL 🌏 JAIN 24 MAHAVEER GANDHI BRAHMAJI NE BRAHM LOK ME EK BRAHM SE HAREK KO SHAKTI PRADAN KI HAI. VO SACH SATYA MADHOOR SUNDAR MITHAS GUNVAN KARTYASHIL DHARM NITH SADAGI SATVA AHAAR VICHAR KARM NITHA OR ATAM SANTOSH SE KUDARAT PREET PRITAM PREM LOVE PYAR SE SAB DETA HAI TAKDIR NASIB GRAHO EK ADDHAR MANA JATA HAI KUDARTI SHAKTI SE SAB MILTA HAI NAHI KE KUTRIM OR BANAVAT SE GALAT JOOT CHODU BANA KE SAB CHALA JATA HAI OR NASHT KHATAM HO JATA HAI. PRUTHVI LOK KA HAREK MANAVI ME EK BRAHM SWAROOP SE SELF ATAM SANTOSH CONFIDENCE ATAM VISVASH ATAL ADAG HONA OR JARURI HAI EXM EK MANAVI KO 6 FEET KI DIVAL KUDANA HAI TO NAHI KUD SAKEGA JAISE KI EK CD HAI USME 1 SE 9 ME SE 10 PAGTHIYA O SE 01 1 SE 09 1 SE 100 1 SE 1000 1 SE 10000 1 SE 100000 ALL UPPAR HI UPPAR SAB BRAHM SHAKTI CHALTA HAI EK BAR AAP CHALNA DOD NA TARANA HAVA ME UDANA TEJ ANDHI ME BAHENA SIKHO AAP APNI COMMENT OR VICHAR MEDIA ME KAHO NEXT FRER POST 📪 ME BAT OR RAZZ KAHUNGA JAY HIND JAIN 24 MAHAVEER GANDHI

Hi! I Just buil opencv 4.8 with cuda on my jetson nano 4gb but it seems that I cannot run those models anymore (need opencv 4.9) Is their a way that I can have access to the 4.8 version model?!

Great

The campaign is closed

Great talk. Thanks Talha & Phil!

Box tv in 2024?..OG

I found the lego robot with depth sensing really interesting (th-cam.com/video/T6AMa_at2N8/w-d-xo.html) Glad you find the little quadruped robot cute when it walks :) (th-cam.com/video/T6AMa_at2N8/w-d-xo.html)

Woa, I made top 20 (Daily Picture Aligner lol)! Thank you for the honor, well done everyone!

wanna see that stufff on a automatic bookscanner

fist comment :-) ever!!

Provide a place where people from the countries which must not be named can contribute to the project. E.g. boosty.

I don't mean to be rude - but the depth quality doesn't seem very high quality at all - much like a Realsense or Oak-D it seems low quality for very coarse applications, able to sense large objects only.

Hello mga hinayupak na BSCS 2A

The same can be said for GPT and other AI systems. Open source drives innovation and invention and marginalizes the divide between those with access and those without.