Find OpenCV Projects on Upwork with Vollna

We're building~

A. Medical social media platform:
1. Patients & hospitals login via email/phone
2. Blog/magazine-style home feed (AI curation)
3. Manual, semi-auto, AI-reporter post creation
4. Auto-post blogs to Facebook, Instagram, Twitter, TikTok, etc. (via API)
5. Patient booking & 1:1 chat
6. QR coupon generation & redemption
7. Web + PWA (Progressive Web App) based

B.Medical photo management app:
1. Electron desktop app (Windows/Mac)
2. Mediapipe web assembly + OpenCV for face/body landmark detection
3. Auto-masking patient photos (front/side/angled)
4. AES256 encryption, local storage, full audit logging

*Needed subscription features for site, App, etc

📌I would give preference to those who can implement the front-end web page nicely.

Please contact me if you can implement it.

We are seeking a skilled professional to calibrate two cameras, focusing on both intrinsic and extrinsic parameters. The project involves setting the point of origin to a specific table. The ideal candidate should have experience in camera calibration techniques and be familiar with relevant software tools. Attention to detail and strong problem-solving skills are essential for ensuring accurate calibration results.

I need an Android developer to finalize a simple app. It should:
Show a homepage with a button to start the camera (Jetpack Compose).

Use the camera to detect glyphs (provided SVGs) with OpenCV.

Call a smart contract on BSC mainnet to get a URL and credentials.

Open the URL in a browser with HTTP authentication.

Test on emulator and deploy on a device like OnePlus 9 5G (Android 14).
Current state: App compiles, camera preview works, glyph detection fails.
Skills: Kotlin, Jetpack Compose, CameraX, OpenCV, Web3j, OkHttp.
Rate: Up to $30/hour, 100% Job Success, 20+ jobs completed, outside USA.
Provided: Zipped project, smart contract details, glyph URLs.

• Include a link to your GitHub profile and/or website

We're building~

A. Medical social media platform:
1. Patients & hospitals login via email/phone
2. Blog/magazine-style home feed (AI curation)
3. Manual, semi-auto, AI-reporter post creation
4. Auto-post blogs to Facebook, Instagram, Twitter, TikTok, etc. (via API)
5. Patient booking & 1:1 chat
6. QR coupon generation & redemption
7. Web + PWA (Progressive Web App) based

B.Medical photo management app:
1. Electron desktop app (Windows/Mac)
2. Mediapipe web assembly + OpenCV for face/body landmark detection
3. Auto-masking patient photos (front/side/angled)
4. AES256 encryption, local storage, full audit logging

*Needed subscription features for site, App, etc

📌I would give preference to those who can implement the front-end web page nicely.

Please contact me if you can implement it.

In the context of a self defense drone detection and jamming system, we'd like to create a system similar to the AARTOS system. We are looking for experts in computer vision, radar, RF, sensor fusion, machine learning and electronics engineering. Any experience with drone detection and jamming is a plus.

We're building~

A. Medical social media platform:
1. Patients & hospitals login via email/phone
2. Blog/magazine-style home feed (AI curation)
3. Manual, semi-auto, AI-reporter post creation
4. Auto-post blogs to Facebook, Instagram, Twitter, TikTok, etc. (via API)
5. Patient booking & 1:1 chat
6. QR coupon generation & redemption
7. Web + PWA (Progressive Web App) based

B.Medical photo management app:
1. Electron desktop app (Windows/Mac)
2. Mediapipe web assembly + OpenCV for face/body landmark detection
3. Auto-masking patient photos (front/side/angled)
4. AES256 encryption, local storage, full audit logging

*Needed subscription features for site, App, etc

📌I would give preference to those who can implement the front-end web page nicely.

Please contact me if you can implement it.

We're building~

A. Medical social media platform:
1. Patients & hospitals login via email/phone
2. Blog/magazine-style home feed (AI curation)
3. Manual, semi-auto, AI-reporter post creation
4. Auto-post blogs to Facebook, Instagram, Twitter, TikTok, etc. (via API)
5. Patient booking & 1:1 chat
6. QR coupon generation & redemption
7. Web + PWA (Progressive Web App) based

B.Medical photo management app:
1. Electron desktop app (Windows/Mac)
2. Mediapipe web assembly + OpenCV for face/body landmark detection
3. Auto-masking patient photos (front/side/angled)
4. AES256 encryption, local storage, full audit logging

*Needed subscription features for site, App, etc

📌I would give preference to those who can implement the front-end web page nicely.

Please contact me if you can implement it.

I'm looking to hire an experienced developer to build a custom automation bot that can streamline my product listing workflow for Depop. The bot should be able to:

✅ Bot Requirements:
Scan Product Links

Accept a list of product URLs.

Visit each link automatically.

Extract Product Data

Download at least 2 images of the product.

Extract the product title and price from the page.

Process Images

Remove the background from each image.

Replace it with a solid yellow background (standardized color).

Save processed images for upload.

Create New Depop Listings

Navigate to the Depop listing page (login credentials will be provided securely).

Upload the edited product images.

Paste in the product title and price.

Fill in required fields (category, condition, etc. — either default or adjustable).

🧠 Preferred Tech Stack (not required, but ideal):
Python + Selenium / Playwright

Image processing via Remove.bg API, PhotoRoom API, or OpenCV

Automation-friendly UI workflows

Experience automating sites like Depop

📦 Deliverables:
Fully working bot with simple interface (command line, GUI, or config file-based).

Instructions for setup and use.

Optional: error logging or skip/retry logic.

这个虚拟产品是一个基于手势追踪的交互式动态视觉特效的核心代码，是用Python和GLSL实现的。

产品核心特点与优势：
1.实时互动与手势追踪： 用户可以通过摄像头进行实时手势追踪，执行脚本会根据手势(五指张开)实时生成并叠加酷炫的动态魔法视觉特效。
2.高质量视觉效果： 特效并非后期模板，而是通过GLSL着色器实时渲染，效果流畅且富有动态美感，类似于电影中的魔法效果(类似奇异博士的某个魔法)。基本不受光线影响，下面有自然光照和基本无光照时的演示图片，视频貌似暂时无法上传，文件大小被限制了。
3.独特卖点： 目前市面上类似产品基本为视频后期处理的特效，这款产品是实时响应的，在直播、短视频创作、线上互动等场景具有很强的吸引力和新颖性。
4.技术栈： Python, OpenCV, MediaPipe (用于手势追踪), ModernGL (用于图形渲染), GLSL (用于特效着色器)。
5.可扩展性强：该Python脚本具备打包成独立可执行软件的潜力，也可以作为核心模块集成到更大的应用中。用户拿到源码后可以根据自己喜好创造和改进，制造出更多”魔法" (若具备相关专业能力)。

我需要的帮助：
我缺乏电商运营和市场推广的经验和精力，因此，我希望能找到一位有经验的电商运营合作伙伴，帮助我将这款产品推向市场，主要工作内容包括但不限于：
1.店铺搭建与优化： 在淘宝或类似电商平台（如抖音小店等，可协商）搭建和装修店铺，使其看起来比较专业且吸引人。
2.商品上架与详情页制作：
撰写吸引人的商品标题和描述文案，突出产品核心卖点。
制作高质量的商品主图、详情页图片和演示GIF/短视频（我可以提供原始的特效演示录屏素材），你也可以录制更好的演示视频，录制起来很方便。
3.基础运营与客服：
设置合理的关键词，进行基础的店铺SEO优化。
处理日常的客户咨询（前期量可能不大），解答用户疑问。
设置虚拟商品自动发货流程（若平台支持）。

我对服务商的期望：
1.有电商平台（最好是淘宝）虚拟产品、软件产品或创意类产品运营经验者优先。
2.对新技术、新奇酷产品有兴趣和理解能力。(主要是这款产品)
3.沟通顺畅，有责任心，能够提供建设性的运营建议。
4.愿意与个人开发者/学生创业项目合作，共同成长。(但也不是很长期且大型的项目)

合作方式与预算：
1.倾向于 “固定服务费 + 销售额提成” 的合作模式，具体比例和金额可详谈。我也愿意考虑其他双方认可的灵活合作方案。
2.这是一个初创项目，初期预算有限，只是初步探索阶段，后期价格可具体商讨，产品具有独特性，希望能找到认可其价值并愿意共同探索的伙伴。
演示录屏文件大小超过了这个限制，动态效果暂时展示不了，感兴趣的朋友请联系我。

Skills: Software Sales, Visual Effects, Operations Management, Interaction Design

**📄 Full Project Specification: Real-Time Learning AI Poker Bot for ClubGG (Private Build)**

---

**🔒 Ownership & Goal**

* This is a 100% privately-owned bot system.
* Goal: Create a real-time learning bot that plays on ClubGG (via LDPlayer) using actual AI—not scripted logic.
* No third-party dependency. Final product must be independent and extendable.

---

**🎯 Phase 1: MVP (Minimal Viable Product)**

**Bot Functions:**

* Launches ClubGG via LDPlayer
* Joins a table (NLH, PLO4/5/6) with assigned name and flag
* Observes the screen using OCR or Computer Vision (NO internal API scraping)
* Detects game state: hole cards, board, stack sizes, position, action history
* Acts via ADB or virtual input (no mouse movement emulation)

**AI Core (Real-Time Learning):**

* Based on Python (TensorFlow/PyTorch or lightweight ML logic)
* Learns behavior of opponents:

* Tags players as Tight, Loose, Aggressive, Passive
* Stores observed actions by player and responds accordingly
* Learns from outcomes (win/loss per action)
* Switches between strategies dynamically: TAG / LAG / Fish / Eco / Random

**Memory System:**

* Player memory (short-term, in-session, and optionally persistent)
* Recognizes players by name and behavior—even if name changes (based on timing and decision patterns)
* Maintains tag/score per player (e.g., Threat Level)

**Multi-Bot Coordination (Initial Layer):**

* Bots on the same table share info via local file or lightweight socket
* Example: If one bot flags a player as “dangerous,” others tighten up

**Control Panel (Local GUI):**

* Assign bot name, country flag
* Select initial strategy (default behavior)
* View known player list and edit memory
* Manually mark a player as “Target”
* Configure join/leave times per bot

**Chat Behavior (Optional but Important):**

* Natural language engine or static preset library
* Sends 1–2 messages per session (humor, trash talk, bait)

---

**🧠 What This Is NOT:**

* This is NOT a click-bot or hotkey macro
* This is NOT a hardcoded flowchart script
* This is NOT a product based on "combat" or "eco" templates as in the screenshots taken from existing bot dashboards (e.g., Fuel, ד, etc.)
* This is NOT a modded ClubGG app

---

**📦 Technical Requirements:**

* Platform: Windows 10/Server 2019
* Emulator: LDPlayer 9 (must support ADB or keyboard input)
* Language: Python 3.10+
* Vision: EasyOCR / Tesseract / OpenCV
* AI Framework: TensorFlow or PyTorch (or reinforcement-lite custom logic)
* Memory: SQLite or local JSON (for modularity and speed)
* Logs: Full game logs per bot with action and decision data

---

**🛡 Security & Integrity:**

* Must avoid detection by ClubGG (use natural input timing, randomized delay)
* No mouse events, use virtual keyboard or ADB
* All models must run locally (no cloud dependencies)

---

**🧪 Phase 2 (Post-MVP / Scaling):**

* Extend to bot farm (multi-instance support)
* Shared central memory between bots
* Advanced coordination logic: target swarm, chip movement simulation
* Cloud dashboard for monitoring 100+ bots
* Bot behavior randomization: IPs, flags, delays, personalities

---

**📬 Delivery Expectations:**

* Working MVP bot + GUI control panel
* Full source code
* Deployment guide (for LDPlayer, emulator setup, bot duplication)
* No third-party calls / SaaS dependencies
* Clean modular code with expansion in mind

---

**💵 Budget:**

* Flexible depending on skill level and commitment
* Payment in USDT (TRC20)
* NDA must be signed

---

**📩 To apply:**
Send:

* Short intro & past experience with AI or poker bots
* Link to prior bot/AI work (GitHub or private demo)
* Cost & delivery estimate for MVP
* Optional: how you’d approach the memory/learning system

---

This is not a toy project. Looking for a real AI partner to build something disruptive, not a script pusher.
Let’s make poker bots smart—for real.

We have an image editing pipeline that we want to automate through AI. The scope of this project is limited to the automation of coloring style. This manually done in Lightroom through presets, but the presets values change based on the images and manual adjustment is needed to ensure the images look good enough. Sometimes, also there are shadows around windows/edges that need to be removed. This should happen while maintaining the resolutions and sharpness of the original images to make sure the image quality doesn't deteriorate. We can discuss things more in the chat.

I need a Python script that mimics Photoshop functionalities. The script should run on MacOS and must have an advanced graphical user interface (GUI).

Core Features:
- Image cropping and resizing
- Color correction and filters
- Layer manipulation

Ideal Skills & Experience:
- Proficiency in Python
- Experience with image processing libraries (e.g., PIL, OpenCV)
- Skills in developing advanced GUIs (e.g., using Tkinter, PyQt, or similar)
- Familiarity with MacOS application development

Please provide a portfolio with relevant projects.

Skills: Java, JavaScript, Python, Software Architecture, C++ Programming

I'm looking for an experienced full-stack developer to build an MVP of a metal roofing calculator with AI-assisted features for dimension and material estimation. This will be the first version of a larger software platform aimed at simplifying roof panel layout, cutting optimization, and pricing for metal fabrication and roofing contractors.
You will be working directly with the project owner (me), who has domain expertise in the metal roofing and purchasing industry.

- Proven experience in full-stack development (Python, JavaScript, React or Vue).
- Experience with AI/computer vision (OpenCV, YOLO, TensorFlow) a big plus.
- Experience working with construction, estimating, or geometry-heavy tools is a bonus.
- Good communication and ability to collaborate on design iterations.
- Comfortable working in sprints or milestones

This project has significant business potential, and if the collaboration goes well, there's long-term opportunity for feature upgrades, AI integration, and mobile expansion.

Looking forward to working with a creative, reliable developer who understands both code and product.

• Describe your recent experience with similar projects
• Include a link to your GitHub profile and/or website
• What frameworks have you worked with?
• Describe your approach to testing and improving QA
• Please list any certifications related to this project

I'm looking for an experienced full-stack developer to build an MVP of a metal roofing calculator with AI-assisted features for dimension and material estimation. This will be the first version of a larger software platform aimed at simplifying roof panel layout, cutting optimization, and pricing for metal fabrication and roofing contractors.
You will be working directly with the project owner (me), who has domain expertise in the metal roofing and purchasing industry.

- Proven experience in full-stack development (Python, JavaScript, React or Vue).
- Experience with AI/computer vision (OpenCV, YOLO, TensorFlow) a big plus.
- Experience working with construction, estimating, or geometry-heavy tools is a bonus.
- Good communication and ability to collaborate on design iterations.
- Comfortable working in sprints or milestones

This project has significant business potential, and if the collaboration goes well, there's long-term opportunity for feature upgrades, AI integration, and mobile expansion.

Looking forward to working with a creative, reliable developer who understands both code and product.

• Describe your recent experience with similar projects
• Include a link to your GitHub profile and/or website
• What frameworks have you worked with?
• Describe your approach to testing and improving QA
• Please list any certifications related to this project

We have a Django application running on ECS Fargate that handles real-time media (video/audio, WebRTC/WebSocket) processing for multiple users. The current setup runs all code (API + heavy video/audio analysis) in a single container. As usage grows, the application is saturating CPU (hitting 100%), causing slowdowns and failures.

Your mission:
- Decouple heavy media analysis (FFmpeg/OpenCV/MediaPipe/Praat/audio) from the main API! These are already in functions

- Refactor the app to use Celery (with Redis or RabbitMQ) for offloading long-running CPU work.

- Deploy dedicated worker containers in Fargate, auto-scaling independently from the API/web app.

- Ensure API/WebSocket container stays responsive
- Build/adjust Dockerfiles as needed, and assist with Fargate/ECS Service/Task configuration.
- Set up (or connect to) Redis (preferably ElastiCache) in AWS as the Celery broker.
- Integrate with our existing stack: Django Channels, S3, and Postgres.
- Implement (or outline) a simple workflow for workers to store results for async fetch/push to clients.


Required experience:
- Django + Celery (async task queues, distributed processing)
- Docker (production-grade), ECS Fargate multi-service deployment
- Redis (and optionally AWS ElastiCache)
- Familiarity with Django Channels/websockets
- Can deliver quickly.

Deliverables:
- Dockerized worker process with Celery consuming and processing jobs
- Refactored Django API/container that enqueues jobs only, and returns promptly
- ECS deployment/config for both services (API and worker), scaling policies/guidance
- Redis setup (locally with Docker Compose, then on AWS - ElastiCache or self-hosted)
- Documentation: how to run locally, how to deploy/scale on AWS


Please respond ASAP with:
Concrete Django + Celery + AWS experience (links/samples preferred)
Your estimate (hours/days) to have a tested solution
Confirmation you can help us scale up, not just “make it work”


This is an urgent migration/scale-up—please only apply if you know production Celery/AWS Docker best practices and are ready to work quickly!

• Are you available to start TODAY?
• Describe your recent experience with similar projects

Will only hire high level profiles. Nothing sent without NDA.

Develop an end‑to‑end “trust chain” that captures, verifies, analyses, and displays images.

3. Core Deliverables (modular)
1. Mobile Capture SDK
• Native Android + iOS wrappers (CameraX / AVFoundation)
• C2PA or equivalent manifest signing (hash, GPS, timestamp)
• Gallery import must be blocked; live‑view capture only
• Lightweight encryption & background upload (TLS 1.3, mTLS)
2. Image Forensics Micro‑service
• Error‑Level Analysis (ELA)
• PRNU sensor noise check
• GAN / deepfake classifier (PyTorch; ready for custom training)
• REST / gRPC endpoint; JSON score output
3. Deduplication & Similarity Search
• Perceptual hash (aHash/pHash/dHash) + FAISS index of CLIP embeddings
• API: query → returns matching IDs + similarity score
4. Vehicle & License‑Plate Recognition
• OpenALPR Community edition integration
• Make/Model classifier (MobileNet v3 or similar)
• VIN OCR (Tesseract or custom CNN)
5. Depth / Damage Estimation
• MiDaS depth map on each image
• Basic metric conversion (wheel diameter reference)
• Output JSON (meta + depth statistics)
6. Optional Physics‑Based “Plausibility” Module
• BeamNG.tech headless simulation via BeamNGpy
• Generate synthetic crash mesh → compare SSIM / LPIPS
7. Back‑End Orchestration
• FastAPI (Python 3.11) micro‑services
• Celery / Redis for async jobs
• Docker‑compose → Kubernetes Helm charts
8. Admin & Investigator Dashboard
• React + Tailwind UI
• Claim table, image viewer with heat‑map overlay, anomaly timeline
• Simple Neo4j graph visual (D3.js) for cross‑claim links
9. Security & Compliance
• Zero‑trust network: mTLS, cert‑pinning, OAuth2 / OIDC (Keycloak)
• Vault‑managed secrets
• Immutable logging (Loki, S3 Object‑Lock)
• Supply‑chain scanning: Trivy, Semgrep, SBOM
10. CI/CD & Documentation
• GitHub Actions (build, test, container sign via Cosign)
• README + Swagger/OpenAPI specs for each service
• Architecture diagram & run‑book

4. Expected Skill Set
• Python (deep‑learning + FastAPI)
• Mobile native (Kotlin, Swift)
• Computer Vision / OpenCV
• Docker & Kubernetes
• React / TypeScript
• DevSecOps fundamentals

5. Confidentiality Note
• Full workflow, datasets, and business context disclosed only after NDA signature.
• No third‑party subcontracting without written approval.
• Code ownership transfers upon final payment.

Skills: PHP, Engineering, Software Architecture, Machine Learning (ML), Deep Learning

We are looking for an AI Developer to help integrate AI-driven features. The ideal candidate should be experienced in machine learning (ML), large language models (LLMs) like ChatGPT, and computer vision to analyze user activity, derive insights from screenshots, automate reporting, and develop intelligent recommendations to enhance productivity.


Responsibilities:
• Develop ML models to classify user activities and measure productivity.
• Implement OCR-based AI models to extract insights from captured screenshots.
• Leverage ChatGPT (LLMs) for AI-driven recommendations and reports.
• Build AI-powered automation for productivity tracking and workflow optimization.
• Integrate AI-driven anomaly detection to flag non-productive or suspicious activities.
• Design and optimize AI models for real-time processing and insights generation.
• Work with the backend team to deploy AI models in a scalable manner.
• Maintain code quality and implement AI testing frameworks.


Requirements
Technical Skills:
• Strong proficiency in Python, TensorFlow, PyTorch, OpenCV, and LangChain.
• Experience with ChatGPT, LLMs, and AI-based conversational interfaces.
• Knowledge of OCR technologies (Tesseract, Google Vision API) and NLP.
• Familiarity with cloud-based AI deployments (AWS, Azure, GCP).
• Experience in data processing, time-series analysis, and feature engineering.
• Strong understanding of REST APIs and AI model integration.
• Experience in real-time AI model optimization and inference.

- Full time opportunity
- Timing : 3PM - 12AM
- Please share CV

1. Project Title
Smart Auto Checkout System – Like Amazon Go (Face Recognition + Item Tracking + Auto Payment)

2. Overview
I want to build an AI-powered smart checkout system for a retail store. The system should identify a user via face recognition at the entrance, detect actions like picking up or placing back items via surveillance cameras, track them in real-time, and charge the user automatically before exit. It should also show the cart in a mobile/web app linked to the user.

3. Key Features
- Face recognition at store entrance

- Action recognition from surveillance cameras (using YOLO or similar)

Reach to shelf

Pick up item

Place back item

Inspect product

- Live cart update for each user

- Payment integration (dummy or real)

- Admin panel to track inventory and sales

4. Deliverables
AI model(s) trained for action recognition from video

Working codebase for face recognition login

Real-time tracking and item list per user

Mobile/web interface to show user cart

Payment module (can be dummy)

Documentation + demo video

5. Tech Stack Preference (if any)
Python (preferred)

Computer vision +LSTM/ TensorFlow / OpenCV

Flask / Django backend

MySQL or Firebase DB

Android/iOS App or ReactJS web app

Skills: Mobile App Development, OpenCV, Tensorflow, Computer Vision, Deep Learning

1. Project Title
Action Recognition Model for Detecting Customer Behavior in Retail Store

2. Overview
I need a deep learning-based action recognition model that can process video input from surveillance cameras and accurately detect customer interactions with store shelves. The model should classify each frame or short sequence into specific actions related to shopping behavior.

3. Target Action Classes
reach_to_shelf – User extends hand toward the shelf

picking_up – User retracts hand from shelf with an item

hand_in_shelf – Hand remains inside shelf (hovering or checking)

inspect_product – Customer inspects item in hand

inspect_shelf – Customer looks at shelf without touching

placing_back – Customer places an item back on the shelf

bold – No person/action visible in frame

4. Requirements
Train or fine-tune a video-based action recognition model (e.g., MSR-RNN, I3D, TSN, TSM, or custom CNN+RNN)

Accepts input video or live stream

Returns class label per frame or clip (windowed prediction)

Should be optimized for retail-like environment (indoor, crowded shelves)

Must include training pipeline + inference script

Accuracy benchmark preferred (confusion matrix, precision/recall)

5. Deliverables
Trained action recognition model (PyTorch/TensorFlow)

Annotated sample dataset (or guide to annotate videos)

Training & inference scripts

Exported model file

Documentation on:

Input format

Model architecture

How to run training & inference

FPS performance metrics

6. Tech Stack (Preferred)
Python

PyTorch or TensorFlow

OpenCV (for video I/O)

Optional: Keras, NumPy, Scikit-learn

Skills: OpenCV, Tensorflow, Keras, Pytorch, Computer Vision

📝 Job Description:
We’re developing a proof-of-concept system that takes a user-uploaded image and text input and programmatically generates:

A print-ready vector file for production (e.g., laser engraving or digital printing)

A photorealistic render of the product with the artwork applied (e.g., on a mug or similar item)

We’re using a pre-designed Illustrator (.ai) template that has text and image placeholder layers. The goal is to build a flexible, headless system that can ingest image/text input and output clean vector + rendered previews without manual design intervention.


🔧 What You’ll Do:

Image Preprocessing

Detect and crop subject (face detection preferred)

Perform basic corrections or orientation adjustment if needed (e.g., turn side-facing photos forward)

Apply stylization filters (posterization, thresholding, or line-art style)

Convert raster image to vector for template placement

Illustrator Automation

Insert the vectorized image into a specified layer of the .ai template

Replace dynamic text fields (e.g., name, description)

Output a high-resolution vector file (.ai or .pdf)

Generate a flattened .png preview

Blender Product Rendering

Take the preview artwork

Apply it to a 3D model of a product (e.g., mug, bottle, plaque)

Configure scene: lighting, background, and camera

Render high-quality mockup image automatically

💻 Tech Stack / Skills Required:

Python (must have)

OpenCV / PIL or similar for image prep

Face frontalization models (e.g., 3DDFA-V2) — we can provide references

Potrace or equivalent raster-to-vector tool

Adobe Illustrator automation (ExtendScript, Node.js, or external control)

Blender scripting and headless rendering via Python API

✅ Deliverables for Proof of Concept:

Vector file (for production: .ai or .pdf with proper layers)

Preview PNG (flattened artwork)

Product Render (Blender .png of 3D model with artwork applied)

You do not need to build a web frontend — we’ll be testing this via Postman or a basic upload form. If this proof of concept is successful, we plan to scale this into a real-time user-facing custom product engine.

💬 To Apply:
Please respond with:

Examples of similar work (image processing, vector automation, product rendering)

Summary of your experience with Blender scripting and Illustrator file manipulation

Confirmation that you're comfortable using or implementing a face alignment tool (like 3DDFA-V2)

Your hourly rate and estimated time to complete the first working version


🎯 Ideal Fit:

Creative technologist comfortable with both 2D and 3D workflows

Able to automate design pipelines for print/e-commerce

Experienced in integrating visual assets into product visuals programmatically

Interested in rapid prototyping and production tooling

Background

Over the past 20 years, our small team (two Ukrainian, one American) has developed an advanced Windows desktop application for the operation of unmanned aircraft and airborne video systems for surveillance.

We expect application and customer requirements and our very long and varied "to-do" list to continue to grow endlessly. We'd like to offload some tasks to another teammember, with the possibility of bringing them on as a permanent part-time or full-time team member.

Our application is mostly in C#, .Net, WPF, with some video-related code in C++ and a few old WinForms displays.

The large code base has grown incrementally / organically over the years and much of it is overdue for refactoring. But our priority is still on rapid forward development as we strive to keep the system competitive and grab market share in a rapidly growing marketplace. Much of the existing code is ugly, but all of the original software authors remain on team and available to assist and to document and clean up existing code as needed,


Initial tasks (fairly easy, Junior level):

Adapt existing code (C#, .NET, GMap) for downloading map tiles from Microsoft Bing Maps to Microsoft Azure. Then integrate additional maps from other similar map tile servers such as USGS.

Convert several old Windows Forms based list displays to WPF and then make numerous additions and refinements to their functionality. These are real-time displays. One provides a filtered view into tens of thousands of log records that accumulate at a rapid rate during operation.
The displays need to provide quick and efficient scrolling to any point in time.



Examples of possible future tasks, requiring more experience:

Refinement of existing DirectShow based system for video display and recording.

Possible integration of various OpenCV based functionality such as object detection and fisheye lens virtual pan/tilt into video stream.

Improvement of integration with 3rd party application via ssl for coordinating with teammembers in the field ("TAK" / "ATAK/CivTAK")


Experience with or interest in drones and related subjects is helpful but not required.

• Describe your recent experience with similar projects

Hi there! Hope you're doing well.

We are a professional guitar studio looking for a freelancer to develop a simple computer vision solution for monitoring our CCTV footage. The goal is to detect if a guitar is removed from its wall mount or storage area without authorisation. We have fixed camera angles inside our studio, and we need a lightweight Python-based script that can monitor these areas and identify when a guitar is no longer visible in its expected spot. Ideally, the system should be able to work either on live CCTV feeds or pre-recorded footage. The output can be basic — even something simple like a console log with a timestamp when a removal is detected would be perfect. You are free to use open-source libraries like OpenCV or YOLO for object detection, but the focus is purely on detecting the absence of guitars, not tracking people or adding complex surveillance features. Experience with computer vision, object detection models, and clean, simple communication is important for this project. We would also like clear documentation included that explains how to set up and run the system. If the initial project goes well, we are open to discussing additional upgrades in the future, such as adding real-time alerts or more advanced monitoring features.

The budget for this project is $150, and the timeline for delivery is 2 to 3 weeks.

1. An LLM annotator (ChatGPT) that generates pseudo labels for given data. They apply prompt engineering using Few-shotting: putting exemplary annotations in the prompt for the LLM (this helps the LLM to annotate).
Moreover, they adopt k-NN example retrieval: sequence embeddings of the text to annotate, and the examples are used to select the examples closest to the new text for few-shot learning.
 They do training cycles, step-by-step training, where first a specific model is trained, new data is annotated by the LLM, and the specific model (BERT) is trained again.
2. An active acquisition mechanism for efficient data selection. They select examples for indirect annotation for which the current model struggles.

3. An automatic reweighting technique to ensure robust learning with noisy labels. They assign learnable weights to the annotated training samples (this makes it possible to reduce the impact of noisy labeling by LLM).

1. Project Title
Smart Auto Checkout System – Like Amazon Go (Face Recognition + Item Tracking + Auto Payment)

2. Overview
I want to build an AI-powered smart checkout system for a retail store. The system should identify a user via face recognition at the entrance, detect actions like picking up or placing back items via surveillance cameras, track them in real-time, and charge the user automatically before exit. It should also show the cart in a mobile/web app linked to the user.

3. Key Features
- Face recognition at store entrance

- Action recognition from surveillance cameras (using YOLO or similar)

Reach to shelf

Pick up item

Place back item

Inspect product

- Live cart update for each user

- Payment integration (dummy or real)

- Admin panel to track inventory and sales

4. Deliverables
AI model(s) trained for action recognition from video

Working codebase for face recognition login

Real-time tracking and item list per user

Mobile/web interface to show user cart

Payment module (can be dummy)

Documentation + demo video

5. Tech Stack Preference (if any)
Python (preferred)

Computer vision +LSTM/ TensorFlow / OpenCV

Flask / Django backend

MySQL or Firebase DB

Android/iOS App or ReactJS web app

We are looking for an AI Developer to help integrate AI-driven features. The ideal candidate should be experienced in machine learning (ML), large language models (LLMs) like ChatGPT, and computer vision to analyze user activity, derive insights from screenshots, automate reporting, and develop intelligent recommendations to enhance productivity.


Responsibilities:
• Develop ML models to classify user activities and measure productivity.
• Implement OCR-based AI models to extract insights from captured screenshots.
• Leverage ChatGPT (LLMs) for AI-driven recommendations and reports.
• Build AI-powered automation for productivity tracking and workflow optimization.
• Integrate AI-driven anomaly detection to flag non-productive or suspicious activities.
• Design and optimize AI models for real-time processing and insights generation.
• Work with the backend team to deploy AI models in a scalable manner.
• Maintain code quality and implement AI testing frameworks.


Requirements
Technical Skills:
• Strong proficiency in Python, TensorFlow, PyTorch, OpenCV, and LangChain.
• Experience with ChatGPT, LLMs, and AI-based conversational interfaces.
• Knowledge of OCR technologies (Tesseract, Google Vision API) and NLP.
• Familiarity with cloud-based AI deployments (AWS, Azure, GCP).
• Experience in data processing, time-series analysis, and feature engineering.
• Strong understanding of REST APIs and AI model integration.
• Experience in real-time AI model optimization and inference.

- Full time opportunity
- Timing : 3PM - 12AM
- Please share CV

Overview
We're building an AI-powered tennis training app paired with a smart ball machine. The app is powered by real-time computer vision and logic-based feedback. We're looking for a skilled iOS engineer or team to help us complete Phase 1 of the app.

This job is contract-based (Phase 1 only), but if things go well, we’re open to continuing with you for Phase 2 and Phase 3.

🔧 What You'll Work On (Phase 1)
1. Fix and improve person detection
- We're currently using Apple's Vision framework for player foot tracking, but the results are inaccurate. You'll help clean this up and ensure we get reliable player positions.

2. Complete the perception → action loop
- Player positions → zone detection → BLE commands sent to the tennis ball machine using hardcoded UUIDs. We have basic logic and lookup tables, but need help finishing and testing this loop.

3. Refactor and organize code
- We need it cleaned up, modularized, and stable enough to run in real-time without overheating or draining the battery.

✅ Optional Add-ons (if time and budget allow)
1. Improve the user experience during court alignment (possibly auto-align with Apple Vision)

2. Add basic difficulty settings (affects command lookup logic)

📱 Tech Stack
Platform: iOS (Swift, SwiftUI)
Frameworks: Vision, CoreML (basic), OpenCV (for homography), CoreBluetooth

Performance goals:
Less than 300ms end-to-end latency (from camera detection to command)
Runs fully offline, heat-safe under 1-hour session

Future Work: If we work well together, you’ll be the first in line for Phases 2 and 3, which include integrating LLMs, pose/ball tracking, and eventually a full custom CV pipeline.

To Apply
Please include:
1. Relevant experience (especially with Vision, CoreML, OpenCV, or BLE)
2. Examples of previous iOS apps you've built with real-time video or camera inputs
3. Your availability and preferred working schedule
4. Whether you're open to continuing for Phase 2+

Too see a full PRD (of phase 1 and briefly for phase 2 and 3), please refer to the attached file.

We are seeking a skilled developer to create a robust application in Python utilizing TensorFlow for computer vision tasks with an industrial GigE camera. The ideal candidate will have experience in integrating camera systems and machine learning algorithms to process and analyze visual data effectively. Your expertise will help us leverage advanced computer vision techniques to enhance our industrial processes. If you have a passion for innovative technology and possess the required skills, we would love to hear from you!

• Please describe your experience with Industrial Cameras (e.g. DALSA Teledyne)
• Describe your recent experience with similar projects

I'm seeking an experienced Python developer to create a real-time face detection application integrated with Streamlit. The application should be able to process webcam video streams and accurately detect faces using OpenCV.

Key Requirements:

• Develop a user-friendly interface with an easy-to-navigate layout using Streamlit

• Implement real-time face detection using OpenCV

• Process webcam video streams

• Ensure high performance and efficiency

Ideal Skills and Experience:

• Proficient in Python

• Experience with Streamlit

• Knowledge of OpenCV and face recognition

• Understanding of AI/ML concepts

• Performance optimization skills

Deliverables:

• Fully functioning Streamlit app

• Complete source code & documentation

• Deployment instructions

Skills: Python, Machine Learning (ML), Artificial Intelligence, Deep Learning

We are looking for an AI Developer to help integrate AI-driven features. The ideal candidate should be experienced in machine learning (ML), large language models (LLMs) like ChatGPT, and computer vision to analyze user activity, derive insights from screenshots, automate reporting, and develop intelligent recommendations to enhance productivity.


Responsibilities:
• Develop ML models to classify user activities and measure productivity.
• Implement OCR-based AI models to extract insights from captured screenshots.
• Leverage ChatGPT (LLMs) for AI-driven recommendations and reports.
• Build AI-powered automation for productivity tracking and workflow optimization.
• Integrate AI-driven anomaly detection to flag non-productive or suspicious activities.
• Design and optimize AI models for real-time processing and insights generation.
• Work with the backend team to deploy AI models in a scalable manner.
• Maintain code quality and implement AI testing frameworks.


Requirements
Technical Skills:
• Strong proficiency in Python, TensorFlow, PyTorch, OpenCV, and LangChain.
• Experience with ChatGPT, LLMs, and AI-based conversational interfaces.
• Knowledge of OCR technologies (Tesseract, Google Vision API) and NLP.
• Familiarity with cloud-based AI deployments (AWS, Azure, GCP).
• Experience in data processing, time-series analysis, and feature engineering.
• Strong understanding of REST APIs and AI model integration.
• Experience in real-time AI model optimization and inference.

- Full time opportunity
- Timing : 3PM - 12AM
- Please share CV

I want to develop a "Gesture Recognition System" with the Intel Real sense camera. It would be very basic, as the system would just react to people stepping on an LED screen, and where they step there must be movement.

• What tools do you use for data mining and visualization?
• Describe your recent experience with similar projects

I am looking to have an expert advisor strategy builder created, using AI to create strategies for currency pairs and gold. This model should be able to generate different strategies either a single strategy or number of strategies combined test them for profitability in the forex market and then be able to export these strategies into metatrader 4 or metatrader 5 expert advisors also giving readable source code.

Core Features of the Strategy Builder
Strategy Generation

Use AI (e.g., genetic algorithms, reinforcement learning, or GPT-based logic) to generate trading strategies.
Strategies can include indicators (e.g., RSI, MACD, Moving Averages), price action, or custom logic.
Backtesting Engine

Test strategies on historical data for currency pairs and gold (XAU/USD).
Evaluate performance metrics: profit factor, drawdown, win rate, Sharpe ratio, etc.
Optimization

Tune parameters using AI to maximize profitability and minimize risk.
Export to MT4/MT5

Generate MQL4 or MQL5 code for the strategies.
Include readable, well-commented source code.
User Interface (Optional)

GUI or web interface to select pairs, timeframes, and strategy types.

AI Techniques to Use
Genetic Algorithms: Evolve strategies over generations.
Reinforcement Learning: Train agents to trade based on reward functions.
GPT-based Strategy Scripting: Use natural language prompts to generate strategy logic.

Output Format
MQL4/MQL5 Files: Ready to compile in MetaEditor.
Readable Source Code: With comments and modular structure.
Performance Report: Summary of backtest results.