Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
    Ferenc Balint-Benczedi
    @sezan92_twitter We actually prefer open communication, so we don't answer the same questions over and over again. It is in the spirit of GSOC to share information. If your question is of general purpose I prefer you ask it here, and only write PMs if it is something very specific (you can also write PMs here)
    Jogender Kota
    hello all, i am new to this community and wish to contribute to Multi-modal Cluttered Scene Analysis in Knowledge Intensive Scenarios.
    I have constructed robots and participated in robotics.
    i have sent an introduction mail content/bremen.de.
    i wish someone could guide me further.
    Jogender Kota
    I have sent an introduction mail at content@uni-bremen.de. *
    Ferenc Balint-Benczedi
    @jogenderkota maybe it is better to send it to me. you'll find my email address in the project description.
    Jogender Kota
    Ok sure.
    Hi all, I have submitted my draft proposal last week, but have not get any feedback so far. Is there someone would like to take a look and guide me further?
    Ferenc Balint-Benczedi
    @johnsonlegend I can not find you in the list of proposals based on your alias here:), could you give me a bit more info?
    @bbferka Sure. I'm Jiongsheng Cai, I have Lisp and Prolog coding experience, and I'm interested in "Plan Library for Autonomous Robots performing Chemical Experiments".
    Ferenc Balint-Benczedi
    @johnsonlegend you should ask my colleague @gheorghelisca , he is the responsible for the planning topics
    Virat Mishra
    @bbferka Sir, i am interested in the cluttered scene analysis project, but i am unable to read the research papers referenced by you. As for the first research paper, it was available on ieeexplore.org but required me to take a paid membership to access it. Upon googling for the second one, only two result links appeared and none of them contained the paper.
    Ferenc Balint-Benczedi
    @viratmishragta4_twitter leave me your email in a PM and tomorrow I can send them to you (I also don't have access from home to them)
    Gheorghe Lisca
    hi @johnsonlegend please write on plan_library_project channel
    Bhairav Mehta
    @bbferka I had a few more questions before I finalize my proposal for the Multi-modal Cluttered Scene Analysis in Knowledge Intensive Scenarios project:
    • If we have background knowledge on the scenario, wouldn't it be beneficial to artifically expand our training set and use machine learning techniques to augment the segmentation algorithms you propose for the project?
    • is the idea of this project to accurately recover the pose of the objects in the scene? Or are we just looking to segment them; maybe they are related questions, but that's why i am not the expert :laughing:
    • what sort of resources would we get in order to work with this project? Would we get just depth clouds, or are there training sets (the paper by Richtsfield et al. uses a hand-annotated dataset for the SVM, would we need to annotate our own data?)
    Bhairav Mehta
    Lastly (sorry for the last minute messages), I understand that we are looking to start from the two algorithms listed in the papers, but they both seem extremely different. Is this a research-based project (which is what i would be more interested in as I prepare for grad school) where we pick and choose the best parts of the papers + other related work, or are you looking for someone to just implement the two algorithms lised in the papers for the robot
    Hello, there! just saying hi! I just found the GSoC but super excited about taking part on it
    Ferenc Balint-Benczedi
    @bhairavmehta95 q1: yes it would, given that we have enough time. As a first step we want to have the algorithms as they are in the framework. q1: accurate segmentation is prefered so far (accurate 6doF might be difficult as specially in the case of symmetrical objects) . q3: We can supply the data in the form of ROS bag files or depth+RGB iamges; and where need(e.g.the SVM from Richtsfield) we can ask authors if they still have the data
    @bhairavmehta95 to your last question: the longer-term goal is to enable the RS framework to handle scenes of increased complexity (clutter in our case). The two papers can also be taken as a guideline on how to do it, but do remember that the scope of GSOC is also to produce usable code. While research in the topic is not just encouraged but IMO a necessary part of this topic, the end result still needs to be a functioning piece of code.
    @bbferka > @viratmishragta4_twitter leave me your email in a PM and tomorrow I can send them to you (I also don't have access from home to them) could you please resend them to me as well? (I am sending my email per PM)
    Ferenc Balint-Benczedi
    Kshitij Goel

    Hi everyone, I am a final year B.Tech student at IIT Kharagpur. I am interested in ROS with PR2 integration in Unreal Engine. I contacted @andreihaidu and he advised me to ask questions regarding my approach here.

    There is a ROS open source implementation (not all functionality, but simple topics, publishers, subscribers and serv-clients) that exists on C#. I have used this before on my Windows system for Unity based on ROS Websockets (roslibjs : https://github.com/RobotWebTools/roslibjs). In my opinion ROS Websockets are the best way to communicate to third party applications from ROS). A JSON parser in C++ would have to be used at the Unreal Engine end, which can easily be written.

    We can use websockets in order to stream data from ROS to Unreal engine and vice-versa. Hence we can control the URDF the way we want from Engine itself.

    @andreihaidu suggested that we link against ros in the linux version of the engine . I would like to discuss more on this. I have previously used WebSockets and have been successful in teleoperation through Unity.

    Jogender Kota
    @bbferka hello, good evening. Just wanted to inform you that I have written a much cleaner proposal and explained what I intend to do in the same.
    Dropped you a mail with the docs link. Apologies for opening a new thread, my bad. Looking forward to your reply. Thank you.
    Archit Singh
    Hello! Intrigued to join the group of like minded people working for a better future!
    Hi everyone! I am a 2nd year B.Tech student and am interested to work on "Multi-modal Cluttered Scene Analysis in Knowledge Intensive Scenarios"
    @bbferka I had certain questions before I submit my final proposal: Would we be provided with training data and approximately how many entries would it contain? Also, can we use any public data set to train our machine learning algorithm? Thanks!
    Ferenc Balint-Benczedi
    @anushree110 if possible we will try to use a public dataset
    Jogender Kota
    @bbferka Hello. I just wanted to inform you that I went through the algorithms and I am getting hold of it. I gained basic concepts of Robosherlock and I am ready to work on it. It would be a great help if you tell me how to proceed further.
    Rajat Kanti Bhattacharjee
    Hi . I am new here and i am pretty clueless at this point. Can anyone suggest me a starting point as to where and what should i be looking into exactly for participating in discussions here
    Mohammed Magdy
    hello everyone
    Hello Everyone I am Very New to this room .From Where Should I start
    Hi,I am interested in AI and have been working on AI and ML for quite some time . I would like to contribute to iai in gsoc 2018,and would like to know how exactly I can be selected. Thank you.
    Hi, I am new here and i am very interested in neural networks part of machine learning. Can that be of any help to my contribution.
    @bbferka Hello, I'm new to the community! I would like to contribute to the project 'Multi-modal Cluttered Scene Analysis in Knowledge Intensive Scenarios' ! Can anyone guide me to go about the same!? Where do i start and what should i learn in order to do the project ?
    Andrei Haidu
    Hi all, we keep in mind your interests for we choose new topics for 2018, however there is no way to tell if we will get accepted again as organization, and how many places we get. As topics in general anything that has to do with AI/ML/Robotics/Vision would be a viable project area. We will update the channel as soon as we get some news about our application (meaning at one point next year)
    Hey, I am new here and want to contribute. Mainly my interest is in reinforcement learning . Is there any project which is related to it.
    Hi everyone !! I am Puneet and would like to contribute to the orgnisation. Can someone guide me where to start? I have a rich experience working on AI/ML projects. Thanks !!
    hi uddeshya here, can any one give a brief overview about how to start contributing in existing codebase?
    Hi all!! Nikhil here. I would like to contribute to the project. Can someone guide me how to start ?
    hello everyone
    got a keen interest in AI and machine learning and would gladly want to contribute
    Any starters for getting into it?
    hi want to get in
    i want to get in
    but have no idea from where to start !!!
    Hello Everybody ! I wish to start contributing. Could you please suggest me projects.
    hello everyone, Shubham here. I would like to contribute. Can someone tell me how to start ?