<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Ranjit SA on Medium]]></title>
        <description><![CDATA[Stories by Ranjit SA on Medium]]></description>
        <link>https://medium.com/@acharya.ranjit?source=rss-63e859dc986b------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Z9Wo0cs4lnFHDMCF2el3mg.jpeg</url>
            <title>Stories by Ranjit SA on Medium</title>
            <link>https://medium.com/@acharya.ranjit?source=rss-63e859dc986b------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 05:05:05 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@acharya.ranjit/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[IGEs to Creo Parametric translation using Deep Learning — Proposal]]></title>
            <link>https://medium.com/@acharya.ranjit/iges-to-creo-parametric-translation-using-deep-learning-proposal-db31740f4093?source=rss-63e859dc986b------2</link>
            <guid isPermaLink="false">https://medium.com/p/db31740f4093</guid>
            <category><![CDATA[translators]]></category>
            <category><![CDATA[cad]]></category>
            <category><![CDATA[creo]]></category>
            <dc:creator><![CDATA[Ranjit SA]]></dc:creator>
            <pubDate>Wed, 13 Mar 2024 10:09:59 GMT</pubDate>
            <atom:updated>2024-03-13T10:24:36.337Z</atom:updated>
            <content:encoded><![CDATA[<h3>IGES to Creo Parametric translation using Deep Learning — Proposal</h3><p>Traditionally Drawing and now 3D model is language of Mechanical Engineer’s Communication. Understandably different 3D CAD software choose to speak differently. They often share information using a neutral format like IGES, STL, DXF etc.</p><p>So a 3D model of one Design software becomes a dumb solid block for other. Although a few translators are available they are not very smart or effective.</p><p>In the age of Data Science art of translation is almost perfected by machines using NLP. Generative AI goes a step further to predict next word/words. Could similar algorithms be used for CAD translation ?</p><p>I believe we could and hence propose to train a Deep Neural Network model like encoders decoders to achieve the same. Being user of Creo (Pro-Engineer in earlier avatar) for years that&#39;s the first CAD tool that comes to my mind. Another reason for using Creo is the ability to translate <strong>Creo part in text file using “Trail File”</strong> and vice-versa <strong>recreating CAD part by playing Trail File.</strong></p><p><strong>Aim of Translator </strong>— Translate IGES to generate Creo part (Trail file) with parametric features.</p><p><strong>Labelled Data </strong>— We will need Labelled data to train the model. We will need thousands and thousands of good Creo Parametric models, their curated Trail File and IGES export.</p><p><strong>Reading IGES in Python </strong>— There are various Python libraries which can read IGES formats like <strong>Pyiges</strong> and <strong>VTK.</strong></p><p><strong>Train Deep Neural Network —</strong>IGES data and corresponding Trail file could be used to train the Neural Network. Architecture could be RNN/LSTM or more complex Transformer. We need to think on the lines of generating captions for image. In our case instead of caption we need to generate Trail file (tall order) for Creo to play.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YJ_uxTnsYFS5cQywf4tlBw.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*s0lUNJ3I20ouxtI31LszGQ.jpeg" /></figure><p><strong>Challenges — </strong>Training such model is a complex task. Apart from computational resources below are some major challenges.</p><ol><li>Need of huge Labelled data. Trail Files will need to be curated to remove unnecessary commands (Pan, Zoom etc.)</li><li>Trail files are not the most stable things and playing them to recreate 3D Model could be frustrating at times.</li><li>Translator will be based on softmax probability models and hence will not be perfect.</li></ol><p><strong>Use Cases- </strong>Although not perfect it could greatly reduce CAD Translation efforts. There could be many use cases for CAD Translators. Below three could be the most common scenarios.</p><ol><li>Acquisitions and mergers of companies which work on different CAD platforms often ends up with challenge of CAD Data migration on one chosen platform.</li><li>Migration of legacy CAD data on new CAD platform is time consuming and often it is done on need basis. Such translators could vastly improve migration productivity. Translation efforts is major hindrance which discourages moving to newer / better platforms.</li><li>Communication with vendors / Customers who are different CAD platforms.</li></ol><p>Who is up for a challenge ?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=db31740f4093" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>