<?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 rohin ramesh on Medium]]></title>
        <description><![CDATA[Stories by rohin ramesh on Medium]]></description>
        <link>https://medium.com/@rohinfablabz?source=rss-8fd5e50d0f1e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*HlNlEVScyiFixSjZQxQW4g.jpeg</url>
            <title>Stories by rohin ramesh on Medium</title>
            <link>https://medium.com/@rohinfablabz?source=rss-8fd5e50d0f1e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 14:25:00 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rohinfablabz/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[A Beginner’s Guide to Mojo Programming: Unleashing Your Creative Power]]></title>
            <link>https://medium.com/@rohinfablabz/a-beginners-guide-to-mojo-programming-unleashing-your-creative-power-1d122024a15a?source=rss-8fd5e50d0f1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/1d122024a15a</guid>
            <dc:creator><![CDATA[rohin ramesh]]></dc:creator>
            <pubDate>Sat, 23 Sep 2023 19:39:48 GMT</pubDate>
            <atom:updated>2023-09-23T19:39:48.582Z</atom:updated>
            <content:encoded><![CDATA[<h4>Introduction:</h4><p>Mojo programming is a term used to describe the art of creating interactive and engaging digital content using a blend of multimedia elements, storytelling, and technology. It’s a dynamic approach that involves creating captivating user experiences to evoke a particular emotional or informational response. In this beginner’s guide, we will explore the basics of mojo programming and how you can start infusing your projects with creativity and engagement.</p><h4>Understanding Mojo Programming:</h4><p>Mojo programming is a fusion of Mobile Journalism (MoJo) and multimedia storytelling. It focuses on leveraging various digital tools and platforms to create compelling narratives through videos, images, audio, and text. The goal is to capture the audience’s attention and convey a message effectively.</p><h4>Components of Mojo Programming:</h4><p>a. Mobile Devices: Mojo often starts with a smartphone or tablet, serving as a versatile tool to capture high-quality photos, videos, and audio.</p><p>b. Apps and Software: Utilize various applications and software for video editing, audio recording, image manipulation, and social media engagement.</p><p>c. Storytelling Skills: Developing a keen sense of storytelling is essential in mojo programming. This involves structuring content in a way that captivates the audience and conveys a clear message.</p><p>d. Visual and Audio Elements: Incorporate engaging visuals, audio effects, and soundtracks to enhance the overall user experience.</p><h4>Key Tips for Mojo Programming:</h4><p>a. Plan Your Content: Start with a clear plan. Outline the message you want to convey and decide on the best way to present it using multimedia elements.</p><p>b. Keep it Short and Engaging: Attention spans are short. Aim for concise and engaging content to maintain the viewer’s interest.</p><p>c. Utilize Social Media: Leverage social platforms to share your mojo content. Utilize features like stories, live streaming, and hashtags to reach a broader audience.</p><p>d. Experiment and Learn: Stay updated with new tools and techniques in mojo programming. Experiment with different styles to find what works best for your target audience.</p><h4>Steps to Get Started:</h4><p>a. Acquire the Right Equipment: Invest in a smartphone with a good camera, a quality microphone, and relevant apps for editing and creating content.</p><p>b. Learn Basic Editing Skills: Familiarize yourself with editing apps like Adobe Premiere Rush, iMovie, or Kinemaster to edit and enhance your videos.</p><p>c. Practice Storytelling: Hone your storytelling skills by creating short narratives and experimenting with different styles.</p><p>d. Engage with the Community: Join mojo programming communities, attend workshops, and connect with fellow enthusiasts to learn and grow.</p><h4>Conclusion:</h4><p>Mojo programming is an exciting journey that combines creativity, technology, and storytelling to produce compelling digital content. With the right tools, skills, and mindset, you can harness the power of mojo programming to captivate your audience and share your message effectively. Start your mojo journey today and unlock the potential to tell stories like never before.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1d122024a15a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Installing mojo programming Language]]></title>
            <link>https://medium.com/@rohinfablabz/installing-mojo-programming-language-286eec4a48a3?source=rss-8fd5e50d0f1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/286eec4a48a3</guid>
            <dc:creator><![CDATA[rohin ramesh]]></dc:creator>
            <pubDate>Sat, 16 Sep 2023 21:56:42 GMT</pubDate>
            <atom:updated>2023-09-16T21:56:42.662Z</atom:updated>
            <content:encoded><![CDATA[<p>Hello Everyone!</p><p>In This Article i gonna teach you how to install Mojo programming language</p><h4>What is Mojo?</h4><p>Mojo is a new programming language that bridges the gap between research and production by combining the best of Python syntax with systems programming and metaprogramming. With Mojo, you can write portable code that’s faster than C and seamlessly inter-op with the Python ecosystem.</p><h4>why mojo?</h4><p>Python has some well-known problems — most obviously, poor low-level performance and CPython implementation details like the global interpreter lock (GIL), which makes Python single-threaded. While there are many active projects underway to improve these challenges, the issues brought by Python go deeper and are particularly impactful in the AI field. Instead of talking about those technical limitations in detail, we’ll talk about their implications here in 2023.</p><h4>Installing in Windows</h4><p>Set up on WindowsNATIVE SUPPORT IS COMING Mojo is not available for Windows yet, but you can use it on a Linux container or remote system. Here’s one way you can do that with WSL2</p><ol><li>Install <a href="https://code.visualstudio.com/">VS Code</a>, the <a href="https://marketplace.visualstudio.com/items?itemName=ms-vscode-remote.remote-wsl">WSL extension</a>, and the <a href="https://marketplace.visualstudio.com/items?itemName=modular-mojotools.vscode-mojo">Mojo extension.</a></li><li>Install <a href="https://apps.microsoft.com/store/detail/ubuntu-22042-lts/9PN20MSR04DW">Ubuntu 22.04 for WSL</a> and open it.</li><li>In the Ubuntu terminal, install the Modular CLI:</li><li>curl https://get.modular.com | \ MODULAR_AUTH=mut_092deaefe9014a22ae4076d8b29324a7 \ sh -</li><li>Command copied to clipboard</li></ol><p>Install the Mojo SDK:</p><p>modular install mojo</p><h4>Installing in macos:</h4><p>Mojo is not available for Apple silicon yet, but you can use it on a remote Linux system. Here’s one way you can do that with GitHub Codespaces:</p><ol><li>Create a fork of our <a href="https://github.com/modularml/mojo/tree/main/examples">GitHub code examples</a>.</li><li>Install <a href="https://code.visualstudio.com/">VS Code</a>, the <a href="https://marketplace.visualstudio.com/items?itemName=modular-mojotools.vscode-mojo">Mojo extension</a>, and the <a href="https://marketplace.visualstudio.com/items?itemName=GitHub.codespaces">GitHub Codespaces extension.</a></li><li>In VS Code, press Cmd+Shift+P and select Codespaces: Create New Codespace. Then select your forked “mojo” repo, and select an instance type with 32 GB RAM.</li><li>Select Terminal &gt; New Terminal, and install the Modular CLI:</li><li>curl https://get.modular.com | \ MODULAR_AUTH=mut_092deaefe9014a22ae4076d8b29324a7 \ sh -</li><li>Install the Mojo SDK:</li><li>modular install mojo</li></ol><h4>Installing in Linux:</h4><ol><li>Verify your computer meets the system requirements.</li><li>Install the Modular CLI:</li><li>curl https://get.modular.com | \ MODULAR_AUTH=mut_092deaefe9014a22ae4076d8b29324a7 \ sh</li><li>Install the Mojo SDK:</li><li>modular install mojo</li><li>Optionally, install the <a href="https://marketplace.visualstudio.com/items?itemName=modular-mojotools.vscode-mojo">Mojo extension for VS Code.</a></li></ol><h4>System Requirements:</h4><ol><li>ubuntu 20.04+</li><li>python = 3.8</li><li>x86–64 CPU</li><li>minimum 4 gb ram</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=286eec4a48a3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Annotating image and audio in label studio]]></title>
            <link>https://medium.com/@rohinfablabz/annotating-image-and-audio-in-label-studio-fd61a3262025?source=rss-8fd5e50d0f1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/fd61a3262025</guid>
            <category><![CDATA[audio]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[signal-processing]]></category>
            <category><![CDATA[label-studio]]></category>
            <category><![CDATA[annotating]]></category>
            <dc:creator><![CDATA[rohin ramesh]]></dc:creator>
            <pubDate>Wed, 10 May 2023 17:12:11 GMT</pubDate>
            <atom:updated>2023-05-10T17:12:11.835Z</atom:updated>
            <content:encoded><![CDATA[<p>Hey folks in this blog I m writing how to use label studio for images labelling and audio labelling and its different features</p><p>INSTALLING LABELSTUDIO: create a new python environment (anaconda or virtual env) and type the below command after updating pip version to latest</p><pre># Install the package<br># into python virtual environment<br>pip install -U label-studio</pre><p>To launch the label studio type:</p><pre># Launch it!<br>label-studio</pre><p>Another way to install label studio is by docker for first you need to install docker. click the link for installing docker <a href="https://docs.docker.com/engine/install/">https://docs.docker.com/engine/install/</a></p><p>after installing docker type the command for installing label studio container</p><pre># Run latest Docker version<br>docker run -it -p 8080:8080 -v `pwd`/mydata:/label-studio/data heartexlabs/label-studio:latest</pre><p>then go to the localhost <a href="http://localhost:8080/">http://localhost:8080/</a></p><blockquote>Steps to annotate images for object Detection</blockquote><p>First open the label studio in localhost then signup. after entering to the projects page click on the create button then you could type the project name and in labelling menu you ll find like…</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*stxKYMr3cZJhTJbgHmB-Lw.png" /></figure><p>here is a bunches of choices you need for your requirements. in this section i ll give a overview of annotating image for object detection and audio classification</p><p>click on “object detection with bounding box section” then you ll need to enter the labels you want for example cat and dog also customise the color for each label. then you ll allowed to import the image data</p><p>after uploading the images go to labelling images</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QoCpXCND-Vb2L3jzilSRCQ.png" /></figure><p>then labelling every datas you could export it in many forms eg: coco, yolo, pascal voc, csv etc……</p><blockquote>Steps to annotate audio datas</blockquote><p>To audio annotation you need to create a new project and type the project name and in data import section upload the audio data which is in .mp3 or .wav file. after enter the labels and click save changes</p><p>after these steps you need to click the label data</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/670/1*52bDLXUu7T84oSw40C35Xw.png" /></figure><p>you will be able to hear and label the data according to your convince after labelling all task you’ll be able to export to many formats</p><p>Thanks for reading and for further project after exporting data please check my other blocks or contact me through rohinr7@outlook.com</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fd61a3262025" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Camera Triangulation for depth and distance analysis]]></title>
            <link>https://medium.com/@rohinfablabz/camera-triangulation-for-depth-and-distance-analysis-6e9da94cc9d7?source=rss-8fd5e50d0f1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/6e9da94cc9d7</guid>
            <dc:creator><![CDATA[rohin ramesh]]></dc:creator>
            <pubDate>Sun, 26 Feb 2023 03:49:59 GMT</pubDate>
            <atom:updated>2023-02-26T03:49:59.770Z</atom:updated>
            <content:encoded><![CDATA[<p>Camera triangulation is a powerful technique for determining the distance and depth of objects in a scene. By using multiple cameras and triangulating the position of the same feature in each image, it is possible to accurately determine the 3D position of the feature and therefore its distance from the cameras.</p><p>To perform distance and depth analysis using camera triangulation, the first step is to calibrate the cameras. This involves determining the intrinsic parameters of each camera, such as focal length and distortion coefficients, as well as the extrinsic parameters, which describe the position and orientation of each camera in the global coordinate system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/380/0*ts5Fa4HbeeA3aO5d.png" /><figcaption>fig 1 ref[1]</figcaption></figure><p>Once the cameras are calibrated, the next step is to capture multiple images of the scene from different positions. These images should overlap and contain enough common features for triangulation to work.</p><p>The next step is to extract and match features in the images. Features can be any distinctive points or patterns that can be easily identified in multiple images, such as corners or edges. Feature detection and matching algorithms, such as SIFT, SURF, or ORB, can be used to automatically detect and match features across images.</p><p>Once the features are matched, the triangulation process can begin. The position of a feature in 3D space can be computed by intersecting the lines of sight from each camera that pass through the feature. This can be done using the Direct Linear Transformation (DLT) or Iterative Linear Triangulation (ILT) method.</p><p>Once the 3D position of each feature is determined, it is possible to compute the distance and depth of objects in the scene. For example, if the cameras are positioned at a known baseline distance from each other, the distance to a feature can be computed by measuring the distance between the two camera positions and the 3D position of the feature.</p><p>Similarly, the depth of objects in the scene can be determined by computing the distance of each feature to the cameras and then projecting the 3D positions onto the image plane of each camera. This results in a depth map of the scene, which can be used for various applications, such as 3D reconstruction, object tracking, and augmented reality.</p><p>In conclusion, camera triangulation is a powerful technique for distance and depth analysis in computer vision and 3D reconstruction applications. By triangulating the position of common features in multiple images, it is possible to accurately determine the 3D position of objects and their distance and depth from the cameras. With the increasing availability of 3D sensors and the development of advanced algorithms, camera triangulation is becoming an increasingly important tool in many fields.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6e9da94cc9d7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Creating a onnx model in python and running inference in C#]]></title>
            <link>https://medium.com/@rohinfablabz/creating-a-onnx-model-in-python-and-running-inference-in-c-98be02d2518a?source=rss-8fd5e50d0f1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/98be02d2518a</guid>
            <dc:creator><![CDATA[rohin ramesh]]></dc:creator>
            <pubDate>Sun, 12 Feb 2023 06:25:26 GMT</pubDate>
            <atom:updated>2023-02-12T06:25:26.464Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>What is onnx?</strong></p><p>ONNX (Open Neural Network Exchange) is an open standard format for representing machine learning models. Converting a model to the ONNX format allows it to be used in a variety of different environments, such as mobile devices or cloud services, without having to recompile the model for each platform.</p><p>converting the pytorch model to onnx</p><p>The process for converting a model to the ONNX format will depend on the framework that the model was created</p><ol><li>First, install the ONNX and ONNX-TF packages: pip install onnx onnx-tf</li><li>Import the PyTorch model and the relevant packages: import torch from torch.autograd import Variable import onnx</li><li>Convert the PyTorch model to an ONNX model: dummy_input = Variable(torch.randn(1, 3, 224, 224)) torch.onnx.export(model, dummy_input, &quot;model.onnx&quot;)</li></ol><p>Now we have converted a image classification pytorch model to onnx format after the conversion we are about to run the model inference using c#</p><p>To perform video inferencing using an ONNX model in C#, you can use the ONNX Runtime library, which provides a high-performance inference engine for ONNX models. Here’s a basic example of how you can use the ONNX Runtime library to perform inferencing on a video stream:</p><p>First, install the Microsoft.ML.OnnxRuntime NuGet package:</p><pre>dotnet add package Microsoft.ML.OnnxRuntime</pre><p>Load the ONNX model into memory:</p><pre>var onnxModel = new OnnxModel(&quot;model.onnx&quot;);</pre><p>Prepare the input data for inferencing:</p><pre>var inputTensor = new DenseTensor&lt;float&gt;(new[] { 1, 3, 720, 1280 }, inputData);</pre><p>Perform inferencing using the ONNX model:</p><pre>var outputTensors = onnxModel.Run(new[] { inputTensor });</pre><p>Extract the results from the output tensors:</p><pre>var outputTensor = outputTensors.First();<br>var result = outputTensor.GetData&lt;float&gt;();</pre><p>This is just a basic example of how to perform video inferencing using an ONNX model in C#. Depending on the complexity of your model and the requirements of your application, you may need to modify this code to suit your needs.</p><p>reference:</p><h3>Object detection with Faster RCNN Deep Learning in C#</h3><p>The sample walks through how to run a pretrained Faster R-CNN object detection ONNX model using the ONNX Runtime C# API.</p><h3>Prerequisites</h3><p>To run this sample, you’ll need the following things:</p><ol><li>Install <a href="https://dotnet.microsoft.com/download/dotnet-core/3.1">.NET Core 3.1</a> or higher for you OS (Mac, Windows or Linux).</li><li>Download the <a href="https://github.com/onnx/models/blob/master/vision/object_detection_segmentation/faster-rcnn/model/FasterRCNN-10.onnx">Faster R-CNN</a> ONNX model to your local system.</li><li>Download <a href="https://onnxruntime.ai/images/demo.jpg">this demo image</a> to test the model. You can also use any image you like.</li></ol><pre>using System;<br>using System.Collections.Generic;<br>using System.IO;<br>using System.Linq;<br>using Microsoft.ML.OnnxRuntime.Tensors;<br>using SixLabors.ImageSharp;<br>using SixLabors.ImageSharp.PixelFormats;<br>using SixLabors.ImageSharp.Processing;<br>using SixLabors.ImageSharp.Drawing.Processing;<br>using SixLabors.Fonts;<br><br>namespace Microsoft.ML.OnnxRuntime.FasterRcnnSample<br>{<br>    class Program<br>    {<br>        public static void Main(string[] args)<br>        {<br>            // Read paths<br>            string modelFilePath = args[0];<br>            string imageFilePath = args[1];<br>            string outImageFilePath = args[2];<br><br>            // Read image<br>            using Image&lt;Rgb24&gt; image = Image.Load&lt;Rgb24&gt;(imageFilePath);<br><br>            // Resize image<br>            float ratio = 800f / Math.Min(image.Width, image.Height);<br>            image.Mutate(x =&gt; x.Resize((int)(ratio * image.Width), (int)(ratio * image.Height)));<br><br>            // Preprocess image<br>            var paddedHeight = (int)(Math.Ceiling(image.Height / 32f) * 32f);<br>            var paddedWidth = (int)(Math.Ceiling(image.Width / 32f) * 32f);<br>            Tensor&lt;float&gt; input = new DenseTensor&lt;float&gt;(new[] { 3, paddedHeight, paddedWidth });<br>            var mean = new[] { 102.9801f, 115.9465f, 122.7717f };<br>            image.ProcessPixelRows(accessor =&gt;<br>            {<br>                for (int y = paddedHeight - accessor.Height; y &lt; accessor.Height; y++)<br>                {<br>                    Span&lt;Rgb24&gt; pixelSpan = accessor.GetRowSpan(y);<br>                    for (int x = paddedWidth - accessor.Width; x &lt; accessor.Width; x++)<br>                    {<br>                        input[0, y, x] = pixelSpan[x].B - mean[0];<br>                        input[1, y, x] = pixelSpan[x].G - mean[1];<br>                        input[2, y, x] = pixelSpan[x].R - mean[2];<br>                    }<br>                }<br>            });<br><br>            // Setup inputs and outputs<br>            var inputs = new List&lt;NamedOnnxValue&gt;<br>            {<br>                NamedOnnxValue.CreateFromTensor(&quot;image&quot;, input)<br>            };<br><br>            // Run inference<br>            using var session = new InferenceSession(modelFilePath);<br>            using IDisposableReadOnlyCollection&lt;DisposableNamedOnnxValue&gt; results = session.Run(inputs);<br><br>            // Postprocess to get predictions<br>            var resultsArray = results.ToArray();<br>            float[] boxes = resultsArray[0].AsEnumerable&lt;float&gt;().ToArray();<br>            long[] labels = resultsArray[1].AsEnumerable&lt;long&gt;().ToArray();<br>            float[] confidences = resultsArray[2].AsEnumerable&lt;float&gt;().ToArray();<br>            var predictions = new List&lt;Prediction&gt;();<br>            var minConfidence = 0.7f;<br>            for (int i = 0; i &lt; boxes.Length - 4; i += 4)<br>            {<br>                var index = i / 4;<br>                if (confidences[index] &gt;= minConfidence)<br>                {<br>                    predictions.Add(new Prediction<br>                    {<br>                        Box = new Box(boxes[i], boxes[i + 1], boxes[i + 2], boxes[i + 3]),<br>                        Label = LabelMap.Labels[labels[index]],<br>                        Confidence = confidences[index]<br>                    });<br>                }<br>            }<br><br>            // Put boxes, labels and confidence on image and save for viewing<br>            using var outputImage = File.OpenWrite(outImageFilePath);<br>            Font font = SystemFonts.CreateFont(&quot;Arial&quot;, 16);<br>            foreach (var p in predictions)<br>            {<br>                image.Mutate(x =&gt;<br>                {<br>                    x.DrawLines(Color.Red, 2f, new PointF[] {<br><br>                        new PointF(p.Box.Xmin, p.Box.Ymin),<br>                        new PointF(p.Box.Xmax, p.Box.Ymin),<br><br>                        new PointF(p.Box.Xmax, p.Box.Ymin),<br>                        new PointF(p.Box.Xmax, p.Box.Ymax),<br><br>                        new PointF(p.Box.Xmax, p.Box.Ymax),<br>                        new PointF(p.Box.Xmin, p.Box.Ymax),<br><br>                        new PointF(p.Box.Xmin, p.Box.Ymax),<br>                        new PointF(p.Box.Xmin, p.Box.Ymin)<br>                    });<br>                    x.DrawText($&quot;{p.Label}, {p.Confidence:0.00}&quot;, font, Color.White, new PointF(p.Box.Xmin, p.Box.Ymin));<br>                });<br>            }<br>            image.SaveAsJpeg(outputImage);<br>        }<br>    }<br>}</pre><p>Thankyou for Reading hope it might help you. if any issues in my writing please mail me in rohinr7@outlook.com</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=98be02d2518a" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>