Generate 870k Realistic Synthetic Images

@juanjosec63

Hi, I have 9 years experience in Unity, Python, C++, image generation pipelines, computer vision datasets, and automated batch processing. For this project, I will build a reproducible Unity-based pipeline that reads your CSV parameters, generates photorealistic semiconductor chip images with accurate colours, dimensions, and shape details, then exports verified JPEG batches with naming, count checks, and logs. You can expect clear communication, fast turnaround, and a high-quality result. I’ve worked on synthetic image generation and ML dataset preparation before, and can structure the pipeline so you can regenerate or tune the full dataset later. Best regards, Juan

@Feriver

Hello I can build a fully automated Unity-based synthetic data generation pipeline that produces your full dataset of ~870,000 photorealistic 2D JPEG images with strict adherence to your CSV-defined parameters for semiconductor chip structures. The system will be architected around a deterministic Unity rendering pipeline where each CSV row directly drives geometry generation, material/color assignment, and camera configuration. This ensures one-to-one reproducibility between input parameters and output images. I will implement a Python orchestration layer to manage batch execution, parallel processing, file naming conventions, and integrity validation to guarantee zero missing IDs and consistent dataset structure. To achieve photorealistic output, I will fine-tune lighting, shading models, and post-processing effects within Unity while ensuring strict adherence to dimensional tolerances defined in your dataset. The pipeline will also include automated validation scripts for random sampling checks against reference images and full logging for dataset traceability and regeneration. Deliverables include the complete Unity project, Python/C++ automation scripts, CSV-driven generation system, dataset validation tools, and fully reproducible pipeline documentation. Thanks, Asif

@OscarMozo

Hi, I can help you with this. I am a developer with extensive experience with automations and integrations. I've helped clients with similar projects. Let me know your interest, Sincerely, Nicolas

@hareshfinadiya

Hi, I am Haresh, having 14+ years of experience in Software Testing Industry. - Having unique blend of knowledge in Quality Product Delivery, Processes Management, Functional testing, Integration and regression testing, load and Perfromance Testing which help me to take the Quality of the software to the next level. - Hands on experience on testing Desktop, Web Based, Mobile application and ERP based application. - Hands on experience on automation testing tools on selenium webdriver, jmeter, katalon studio, Appium, cypress, selenium with TestNG freamwork etc.. - Thorough understanding of Product Delivery Life Cycle, Software Testing Life Cycle and Software Development Life Cycle. - Experience in Well conversant with writing Test plan,Test Cases,Bug report, Release Note and Product Health Report. - Worked in various domains like Finance, Retail, Web Portals, Healthcare, ecommnerce, CMS, Eduction Portal, Life Insurance, ERP system etc. - I do have require mobile devices to test mobile view or applications like android and iOS applications. - I have hands on experience with Git, postman, MSSQL Server. Kindly review my profile and let me know you view over the same. Thanks, Haresh

@seandinwiddie

✨✨✨ ✨✨✨ ✨✨✨ ✨✨✨✨✨✨ ✨✨✨ ✨✨✨ ✨✨✨✨✨ ⏳ Timeline : 7 days ✅ Budget : 460 USD Hi, Dear. Portfolio : https://www.freelancer.com/u/seandinwiddie I can build your Unity-based synthetic dataset generation pipeline to reliably produce ~870,000 photorealistic semiconductor-style 2D JPEG images driven directly from your CSV specifications, with full reproducibility and strict parameter compliance. I will start by designing the Unity scene and asset system to accurately represent chip geometry, lighting consistency, and camera framing aligned with your reference dataset. Next, I will implement a Python/C++ orchestration layer to parse your CSV, control batch rendering, enforce naming conventions, and validate that every generated image matches its parameter row (color, dimensions, and positional constraints). Finally, I will optimize batch rendering, logging, and verification scripts to ensure zero missing IDs, consistent resolution, and automated QA sampling (including 500-image random inspection tooling). The system will be fully reproducible so you can regenerate, extend, or modify the dataset later without rebuilding the pipeline. I will also ensure performance optimization for large-scale rendering runs (queue batching, GPU utilization, and disk I/O control). Looking forward to collaborating. Best Regards. Sean D.

@mengw3

With over a decade's worth of experience, including simulations using advanced 3D modelling techniques, I am a perfect fit for your project. Having participated in the AirPlay 2 Integration on OpenWRT contest and emerging as the winner, I have hands-on experience working with large datasets, ensuring excellent reproducibility and strict adherence to defined parameters. Leveraging my expertise in Unity, Python, and C++, I will build an efficient automated pipeline that thrives on providing quality batch generation with zero missing IDs. My extensive knowledge of managing data processing and verification through scripts will guarantee consistent resolution throughout all 870k images. Moreover, my proficiency in AI and ML will be indispensable when it comes to sharply rendering realistic 2-D images under dimensional tolerances. Ultimately, partnering with me means securing efficient problem-solving skills alongside the value of my 20 years of proven experience. Let's talk prospectively to establish an incredible timeline through which we can exceed your project goals together!

@souvikbasu

Semiconductor chip imaging pipelines have one hidden failure mode: colour drift across 870k images when the CSV spec has tight tolerances. Here's how I'd prevent it: • Unity 2D scene with a CSV-reader component mapping each row to exact colour, X-Y position, and shape before every render • Python batch manager handling naming conventions, job chunking, and a count-verification log after every 10k images • Automated spot-check script sampling 500 random IDs, overlaying CSV references, and flagging deviations — meeting your acceptance criteria by design • Reproducible project files with a single CSV-swap to regenerate or tweak the entire dataset I also saw the clarification thread — the hybrid approach (small Unity-rendered base + photorealistic augmentations) is exactly how I'd structure this to hit 870k efficiently without sacrificing visual realism on spot checks. Two questions: Are you using orthographic or perspective projection in Unity? And do the JPEGs need a specific DPI beyond what the CSV specifies?