Find OpenCV Projects on Upwork with Vollna

We're looking for an experienced Computer Vision Engineer / Python Developer to build a lightweight, scalable people counting system that uses YOLOv5 or YOLOv8 and OpenCV to monitor foot traffic through existing store cameras (RTSP feeds). (Multiple store locations and companies)

This project also includes building a command-line tool (CLI) that allows our internal team to easily configure new camera feeds, set detection zones, view counts, and export data — We have a web dashboard to incorporate the data.

What We Need Built:
Core System
Detect people using YOLOv5 or YOLOv8 on live or recorded video

Line Count

Zone counts/ time./ traffic

Ingest RTSP/IP camera streams via OpenCV

Implement line-crossing or zone-based counting logic

Handle multiple feeds efficiently

Output counts in real-time or batch (e.g., hourly, daily) to:

CSV

SQLite (optional)

CLI Tool
A clean command-line interface that allows us to:

Add/remove/edit camera feeds
counttrack add-camera --name "Front Door" --url "rtsp://..."

Define detection zones (e.g., crossing lines) via coordinates
counttrack set-zone --camera "Front Door" --coords "[(100,200),(300,200)]"

Start/stop processing on specific feeds
counttrack start --camera "Front Door"

Export data or view live counts
counttrack export --date today --camera "All"
counttrack status --camera "Front Door"

Nice-to-Haves (Optional for Phase 1):
CLI preview feature: show video frame for drawing detection zones

Multi-zone or multi-direction logic

Ability to run on Raspberry Pi or Jetson Nano

Optional web-based dashboard in future phases

Ideal Experience:
YOLOv5 / YOLOv8 (Ultralytics)

Python & OpenCV

Experience with RTSP / IP cameras

CLI frameworks like Typer, Click, or argparse

Data handling: CSV, SQLite, or simple JSON configs

(Bonus) Experience running CV on edge devices

Deliverables:
Fully functional people-counting system for store camera feeds

Python-based CLI tool with documentation

Config files / sample zones

Setup instructions (can be pip-installed or Dockerized if preferred)

All source code and rights transferred

Budget:
$1,500 – $3,500 USD

Open to hourly or fixed-price bids depending on timeline and experience

To Apply:
Please include:

Links to similar projects (GitHub, demos, etc.)

Your experience with object detection and CLI tools

A short summary of how you'd approach this project

Your availability and estimated delivery time

Skills: Computer Vision, OpenCV, Python, Machine Learning, Raspberry Pi

Offline Face Recognition and Anti-Spoofing Implementation for Mobile App (Flutter/Dart)

Description:
We are looking for an experienced Flutter developer with expertise in machine learning and face recognition to implement an offline solution for face recognition and anti-spoofing in our mobile app. The solution should work seamlessly without relying on internet connectivity and provide robust spoofing detection.

Requirements:

Implement face recognition using TensorFlow Lite, OpenCV, or an alternative SDK.

Integrate anti-spoofing functionality using liveness detection or advanced algorithms.

Offline operation without relying on cloud-based APIs.

Flutter mobile app integration for both Android and iOS.

Ensure accuracy and speed of face recognition on mobile devices.

Provide documentation and guidance for future maintenance.

Deliverables:

A Flutter-based application with integrated offline face recognition and anti-spoofing features.

Model(s) or SDK(s) integrated for offline functionality.

Documentation on how to use and update the face recognition system.

Skills: Smartphone, Tablet, AI Mobile App Development, Android App Development, iOS, Microsoft Windows, Android, Flutter, iOS Development

I need an experienced developer with a strong background in image processing and computer vision. The tasks involve apply basic image processing techniques to a grayscale image focusing on image transformation, pose estimation, and ROI identification.

Key features
- setup Translation, Rotation, and Scaling metrics.
- Conduct 2D pose estimation tasks for accurate analysis of the image
- Use edge detection techniques to identify regions of interest (ROI) in grey scale image
- Apply Beer Lambert Law for necessary image processing requirements.

Ideal Skills:
- Proficiency in Opencv and python.
- Strong understanding of image registration, edge detection, and pose estimation.
- Experience with edge detection techniques and the application of the Beer Lambert Law.
- Familiarity with breast tissue mimic medical imaging is a plus.

Skills: Python, Testing / QA, Image Processing, OpenCV

We’re building an AI-powered strength coaching SaaS platform that automates program design, adapts training based on real-time data, and enhances how coaches manage athletes.

If you have experience in AI/ML development, fitness-related applications, and backend API integration, we’d love to work with you. This is an opportunity to develop a cutting-edge tool that helps strength coaches, trainers, and physical therapists optimize training programs with AI-driven recommendations.

We need an AI/ML engineer who can develop the machine learning models that drive intelligent strength training recommendations and automate program design based on user inputs.

The system should:

✅ Auto-generate strength training programs based on user profile data (age, goals, injury history, experience level, etc.)
✅ Adapt workouts dynamically based on real-time feedback and performance tracking
✅ Store and process user training history for progress analysis & AI adjustments
✅ Integrate via API into a web or mobile application (React.js or React Native/Flutter preferred)
✅ Potential for computer vision-based movement screening (future feature expansion)

This is a long-term project, but the first goal is building an MVP within 3-6 months.

We expect you to:
✔️ Develop the AI/ML model for personalized program design (Python, TensorFlow, PyTorch, or Scikit-learn)
✔️ Set up a backend API to allow our app to communicate with the AI (FastAPI/Django preferred)
✔️ Deploy the model on AWS/GCP/Azure for seamless access
✔️ Provide clear documentation for how the model works, how it can be improved, and how data is processed
✔️ Collaborate with our frontend team (or assist in finding one) to ensure smooth integration

Bonus points if you have experience with:
➕ Computer vision (OpenCV, MediaPipe) for movement screening
➕ Wearable API integrations (Whoop, Apple Health, Fitbit, etc.)

💡 Machine Learning & Deep Learning (Supervised learning, reinforcement learning, NLP is a plus)
💡 Python (TensorFlow, PyTorch, Scikit-learn)
💡 Backend Development (FastAPI, Flask, or Django for API development)
💡 Database Management (PostgreSQL, MongoDB, or Firebase)
💡 Cloud Deployment (AWS, GCP, or Azure)
💡 API Design & Integration (for frontend connectivity)

We can provide the data on:
⭐ Experience in fitness, sports science, or training analytics
⭐ Strong understanding of periodization & strength training principles

We value clear and consistent communication to ensure a smooth workflow. You’ll work closely with me, the founder, and a potential frontend developer.

🔹 Weekly check-ins via Slack or Zoom
🔹 Use Trello/Asana for task tracking
🔹 Share updates & challenges proactively

We’re looking for someone who can bring ideas to the table and suggest ways to improve our AI system—not just follow instructions.

Skills: AI Model Integration, Database Management, Machine Learning, Deep Learning, Python, Data Engineering

I have MAUI application, I am looking for somebody to help build a face recognition engine within the MAUI mobile application targeting iOS and Android. I have someone who has done something like this using MAUI before with proven success rate. Note that the targeting people are kids between age of 7 to 16. The engine needs to be able to support kids. Show me example in video to showcase your capability. I need guarantee that it will work.

What already available, a list of people, 4 photos per person in jpg/png.

What you need to do:
The engine should take in the list of people with 4 photos each, do the necessary pre-processing, such as remove unwanted background, cropping etc. Train the model and then fire up the camera with camera view ready for face recognition. For every faces that is detected from the list, store the ID in a list. Once the "END" button is click, return all the success ID back to the calling function. Check the attached video for some idea, improvement is needed on the engine.

Condition:
1) Young kids between 7 to 16 years old
2) Support with or without glasses on
3) Good to have - with mask on

More details will be provided during the evaluation process.

Skills: Face Detection, Machine Learning, Computer Vision, Neural Network, TensorFlow, OpenCV, Image Processing, Xamarin

I’m looking for a Python and Excel expert for automating web data scraping from e-commerce sites. The primary focus of the project is scraping product details from various e-commerce platforms.

Key Responsibilities:
- Develop a script in Python for scraping product details from e-commerce sites.
- Automate the data collection process to streamline the workflow.
- Organize the scraped data in Excel for easy analysis.

Ideal Skills:
- Proficiency in Python and Excel.
- Experience in web scraping, specifically from e-commerce sites.
- Ability to automate processes and improve efficiency.
- Strong data organization skills.

Please provide examples of similar projects you've completed in your proposal.

Skills: Python, Excel, Web Scraping, Data Mining, OpenCV

More info will be discussed in private
More info will be discussed in private
More info will be discussed in private
More info will be discussed in private
More info will be discussed in private
More info will be discussed in private
More info will be discussed in private
More info will be discussed in private
More info will be discussed in private
More info will be discussed in private

Skills: Python, Python Script, FastAPI, OCR Software, OCR Algorithm, Tesseract OCR, Computer Vision, Image Annotation, Roboflow, CVAT, GPT-4, YOLO, Machine Learning, Database, OpenCV

Experienced Developer Needed for Image Processing Project using OpenCV

Description:
We are seeking a highly skilled developer with expertise in image processing and machine learning to help build a sophisticated analysis tool for urine test strips. This project involves developing a mobile application that can interpret chemical test results from photographed urine strips using OpenCV and potentially integrating machine learning techniques for enhanced accuracy.

Responsibilities:
- Develop and optimize image processing algorithms using OpenCV to analyze urine test strip images.
- Implement preprocessing steps to enhance image quality for analysis.
- Design and train machine learning models if necessary to classify and interpret test results.
- Work collaboratively with our team to integrate the image processing backend with a user-friendly mobile interface.
- Ensure the application adheres to data privacy and security guidelines relevant to health data.

Requirements:
- Strong proficiency in Python and/or C++, with a solid background in using OpenCV for image processing tasks.
- Experience in building and deploying machine learning models, with familiarity in frameworks like TensorFlow or PyTorch (a plus).
- Previous experience in developing health-related applications, particularly those involving image analysis (highly desirable).
- Excellent problem-solving skills and the ability to work independently or as part of a team.
- Good communication skills and proficiency in English.

Project Duration:
- 3-6 months, with the possibility of extension based on project needs and performance.

Skills: OpenCV, Python, C++, Tesseract OCR, Computer Vision, Machine Learning, Artificial Neural Network, Deep Neural Network, Mobile App Development, Microsoft Visual C++

We are seeking a Python developer to create a script that accurately detects cat and dog poses. The solution should be efficient and avoid using complex libraries such as mmmpose or YOLO. It must be compatible with macOS M2 to ensure optimal performance. If you have experience in building similar applications and can provide a reliable solution, we want to hear from you. Please include examples of previous work related to computer vision or pose detection in your proposal.

Skills: Python, Machine Learning, TensorFlow, OpenCV, Python Script

I'm looking for a creative and technically skilled developer to build a custom animation pipeline that takes a static cartoon-style face image (think Studio Ghibli aesthetic) and an MP3 audio file, and outputs a lip-synced video with captions.

Unlike traditional deepfake or avatar platforms, this tool will be designed specifically for stylized 2D characters — not real human faces — and should support full-length audio with no time limits or platform restrictions.

Key Features:
✅ Accepts a front-facing, cartoon-style face image (static)
✅ Accepts MP3 audio input
✅ Transcribes the audio to generate accurate captions
✅ Maps phonemes to mouth shapes (visemes)
✅ Overlays stylized mouth frames onto the image to simulate lip movement
✅ Outputs high-quality animated video (MP4) with synced subtitles
✅ Designed for unlimited-length audio (not capped at 1–2 minutes)

Preferred Technical Approach:
Phoneme/viseme extraction via Whisper or similar STT model

Animation via sprite-based mouth overlays or 2D morphing

Video assembly using MoviePy, OpenCV, or FFmpeg

Captions styled and burned into the video using timestamped transcriptions

No reliance on tools like Wav2Lip or deepfake models

Bonus (Nice to Have):
Experience with animation pipelines, 2D rigging, or Unity/Unreal

Familiarity with anime/cartoon aesthetics

Ability to build a simple interface (Gradio, Streamlit, or local script)

Modular code for future expansion (e.g., head motion, blinking, expression changes)

Deliverables:
A functional script or app that processes one image + one MP3 → outputs video

Support for unlimited audio length (within resource limits)

Stylized caption overlay

Setup and usage documentation

Timeline & Budget:
Open to hourly or fixed price proposals

MVP delivery preferred within 2–3 weeks

To Apply:
Please include:

Relevant experience with audio-visual or AI animation projects

Examples of similar work (cartoon, lip sync, or animation tools)

A brief outline of how you'd approach this project

Skills: Python, AI Development, Video Processing, Machine Learning, OpenCV, Automatic Speech Recognition, Cartoon Animation

Development of an AI-Powered Fitness Buddy – Software & AI Integration

2. Project Overview
The AI-Powered Fitness Buddy is an automated fitness assessment system designed to scan a person’s body, analyze their physical attributes, and provide fitness and diet recommendations based on AI-driven analytics.

🚀 Important Note:

I will assemble and set up the hardware.

The freelancer will be responsible for all programming, AI processing, and software integration remotely.

3. Expected Functionality
✅ Step 1: User Entry & Interaction

System activates when a person enters the scanning cabinet.

User selects a fitness goal via voice or touchscreen.

✅ Step 2: Body Scan & Data Collection

The system collects height, weight, body fat %, and other metrics via sensors.

✅ Step 3: AI-Based Fitness & Diet Suggestion

AI processes data and provides personalized fitness & diet plans.

✅ Step 4: Progress Tracking

User data is stored, and progress is tracked over multiple visits.

4. Scope of Work (Software Only)
The freelancer will handle:
🔹 Sensor Data Processing – Writing code to interpret data from height, weight, and body composition sensors.
🔹 AI & Machine Learning – Implementing algorithms for BMI calculation, fat percentage estimation, and fitness recommendations.
🔹 Voice & Touch Interface Development – Enabling user interaction through speech recognition and a touchscreen UI.
🔹 Progress Tracking System – Storing user data and comparing past vs. present readings.
🔹 Remote Debugging & Support – Assisting with software integration and troubleshooting.

❌ No Hardware Assembly Required – I (the client) will handle all physical setup and sensor connections.

5. Required Freelancer Skills
AI & Machine Learning (TensorFlow/PyTorch)

Embedded Systems Programming (Python, Raspberry Pi)

Computer Vision (OpenCV for body scanning)

GUI Development (Tkinter/PyQt for touchscreen UI)

Speech Recognition APIs (Google/Azure for voice commands)

Database Management (SQLite/PostgreSQL for user data storage)

6. Hardware Setup (Provided by Client)
The freelancer will write software for the following hardware components:

Component Purpose
Raspberry Pi 4 / Jetson Nano AI Processing & System Control
Load Cell Sensor Weight Measurement
LiDAR / Ultrasonic Sensor Height Measurement
Infrared Camera / Thermal Sensor Body Fat Analysis
MPU6050 / IMU Sensor Posture & Balance Analysis
Touchscreen User Interaction
Microphone & Speaker Voice Command Recognition
📌 How We Will Work Together:

I will connect the sensors & provide live test data.

You will write the software & provide test scripts.

We will troubleshoot together via remote debugging.

7. Project Milestones & Payment
🔹 Milestone 1: Sensor Data Processing (30% Payment)

Writing Python code for all sensors

Validating real-time sensor readings

🔹 Milestone 2: AI Fitness Recommendation System (30% Payment)

Implementing AI-based fitness & diet suggestions

Testing accuracy of recommendations

🔹 Milestone 3: GUI & Voice Interaction (40% Payment)

Developing a smooth user experience

Integrating all modules & final debugging

8. Next Steps
📌 How to Apply:

Share examples of your previous work with AI, Raspberry Pi, and GUI development.

Explain your approach to remote debugging & software delivery.

Provide your estimated timeline & pricing.

💡 Looking for a skilled freelancer who can work remotely & ensure smooth software integration!

Description:

We are seeking a Python developer to assist with a tennis ball tracking project, where the key challenge is accurately detecting and tracking small objects (tennis balls) in real time across various video recordings.

What We Need:

· A Python developer who’s comfortable with OpenCV and Ultralytics YOLO (or similar) to build or fine-tune a system that can handle these fast, small objects in real time.
· Someone who gets the quirks of small-object detection – they’re easier to miss, so we need creative solutions (higher-res input, tiling, specialized training, etc.).
· A solution that’s robust across several videos (different angles, lighting, speeds, etc.).
· Well-documented code: This means a clear README with detailed steps, listing all dependencies, explaining the techniques used, and including a demo video showing the system in action.

You’ll Be Doing Stuff Like:

· Tweaking existing models (e.g., YOLOv8) or training new ones so they can spot these tiny yellow fuzzballs.
· Making sure the system runs fast enough for real-time or near-real-time tracking.
· Handling the usual suspects: occlusions, motion blur, partial visibility, etc.

If this sounds like your thing, let us know! We’re excited to see your approach for handling small, speedy tennis balls and look forward to a final product that’s well documented and easy to replicate.

Here an example tennis macthes where we will run your code: https://www.youtube.com/watch?v=x1TQKedRfgA&t=248s

Thanks!!

Skills: Computer Vision, Deep Learning, Machine Learning, Object Tracking, Python

I’m looking for a developer to create an algorithm that integrates with my app’s existing technology. The app scans a person’s face and generates one of six lash maps, which act as a guide for where each lash should be placed.

The algorithm should:
• Read the lash map and determine the correct type of lashes to use (individual, fans, or spikes) based on the selected set.
• Place the lashes in the correct position (inner, outer, middle, or straight) with the correct length and curl.
• Work in real-time as an augmented reality (AR) filter, ensuring that as the user moves their head, the lashes stay in place and adjust naturally—similar to a Snapchat filter.

For example, if the lash map specifies a volume set with C curl and an 8mm length in the inner corner, the algorithm should:
• Select only fans (since it’s a volume set).
• Use C curl (as defined by the map).
• Place 8mm fans in the inner corner using the correct perspective.

This app is designed for eyelash technicians who specialize in beauty and eyelash extensions. The system has already been programmed to measure a person’s eyes and automatically generate the ideal lash map with the correct curl, lengths, and set based on those measurements. The app already knows what lashes to use—users do not input this information manually.

Additionally, you will not need to create any lash images or assets—I already have them ready to be used. Your job is strictly to integrate this existing system into the real-time lash try-on feature, allowing users to see the lashes applied to their eyes and moving naturally as they turn their head.

You don’t need prior knowledge of eyelash extensions—everything related to how lash sets work will be explained in a Google Doc. You’ll also be able to work with my developers to understand how the system generates lash mapping.

This project requires expertise in:
• Algorithm development for automated lash placement.
• Computer vision & facial tracking to ensure the lashes move naturally with the user’s face.
• AR development to create a real-time filter experience.

My current developers have tried to do this but they can’t seem to train a machine learning model that can properly place the individual lashes according to the given lash map. If you have experience with facial tracking, real-time AR filters, and algorithm development, I’d love to hear from you!

Skills: Machine Learning, Computer Vision, Augmented Reality, iOS, Android, OpenCV, Artificial Intelligence

I'm conducting a research project focused on AI-based sports video analysis, specifically soccer. The objective is to classify players into two teams and other roles (e.g., referees, goalkeepers) using unsupervised learning techniques. The outcome will be a comparative analysis paper evaluating different approaches.

I'm looking for support with both the technical implementation in Python and the preparation of the research paper in LaTeX.

Scope of Work:

- Implement object detection using models such as YOLO, Faster R-CNN, or SSD

- Apply clustering techniques (e.g., K-Means, GMM, color-based segmentation) to classify players based on jersey colors

- Evaluate and compare the performance of these approaches

- Contribute to writing the research paper using LaTeX, with a focus on methodology, experiments, results, and discussion

Requirements:

- Proficiency in Python and relevant libraries (OpenCV, scikit-learn, PyTorch or TensorFlow)

- Experience with object detection and unsupervised learning methods

- Familiarity with academic writing and structuring papers in LaTeX

- Ability to explain and analyze experimental results clearly

This project is ideal for someone with a background in computer vision, machine learning, and academic research. If you're interested, please share your relevant experience and any sample work if available.

Skills: Machine Learning, Natural Language Processing, Artificial Intelligence, Python

I am seeking an experienced OCR (Optical Character Recognition) and image processing expert to extract a license plate number from surveillance footage. The vehicle in question is a gold/tan Ford Escape with a red license plate featuring white numbers.

Two video files and a photo have been uploaded to this job post. Your task is to review them and enhance any frames necessary to extract license plate information. Even a partial plate may be sufficient.

IMPORTANT:
Auto-generated job replies created by ChatGPT or similar tools will be automatically disregarded.

By applying, you agree to forfeit payment if you cannot successfully complete the job.

Do not apply if you are unsure about your ability to complete the task.

Please review both videos carefully before submitting your proposal.

Requirements:

Deep experience with OCR using tools like Tesseract, EasyOCR, Google Vision, or YOLO.

Strong background in image and video enhancement, including frame-by-frame analysis, super-resolution, denoising, and sharpening.

Ability to handle low-light and motion-blurred footage.
Familiarity with automated license plate recognition (ALPR) tools and forensic image analysis.

Fluency in Python, OpenCV, or equivalent frameworks.

Deliverables:

Enhanced image(s) or video frames clearly showing the license plate area.

OCR output with extracted characters (partial plate acceptable).

Brief explanation of the techniques and tools used.

Timeline:

Completion expected within 48–72 hours of contract start.

Application Instructions:
Do not apply unless you have reviewed the two attached videos and image.

In your proposal, explain your approach to this task and reference similar past projects.

Confirm you understand and accept the condition of forfeiting payment if unsuccessful.

Skills: Computer Vision, Image Processing

I need a crowd-counting solution for mega-events (e.g., Carnival, concerts with 200K+ attendees) using drones and AI. The goal is to validate an MVP combining:
- Drone-captured video feeds
- Computer Vision APIs (Google Vision, YOLOv8, Crowd Counting models, or your suggestions)
- Real-time dashboard for density analysis and safety monitoring.

**Technical Requirements:**
1. Proficiency in **Python** and computer vision frameworks (OpenCV, YOLOv8, TensorFlow/PyTorch).
2. Experience integrating AI/ML APIs (e.g., Google Vision, Roboflow) and video/image processing.
3. Familiarity with **drone video capture** (DJI SDKs, flight planning) and edge computing (optional but a plus).
4. Ability to build a lightweight dashboard (Streamlit, Power BI, or similar) for real-time data visualization.

**Expected Deliverables:**
✅ System to process drone footage and generate **people count + density heatmaps**.
✅ Integration with at least **1 computer vision API** (e.g., Google Vision).
✅ Technical documentation + validation report from a test event.

**Why Join This Project?**
- **Real-world impact:** Your solution will be tested at Brazil’s largest events.
- **Long-term potential:** Opportunity for ongoing collaboration to scale the product.

**How to Apply:**
1. Share your portfolio (GitHub, past AI/computer vision projects).
2. Briefly outline your approach (e.g., API selection, ML model design).
3. Specify your availability and estimated timeline.

Skills: Python, OpenCV, Tensorflow, Computer Vision, YOLO

We are a logistics company looking to streamline and automate our document management process. We need a skilled AI/ML Developer to develop an MVP that intelligently processes, organizes, and retrieves logistics-related documents using AI-driven automation.

Responsibilities:

Design and develop an AI-powered document management system for handling invoices, bills of lading, shipping manifests, and contracts.

Implement OCR (Optical Character Recognition) and Natural Language Processing (NLP) for automated document classification, extraction, and tagging.

Build a searchable database with efficient indexing and retrieval capabilities.

Ensure data security and compliance with industry regulations.

Develop APIs for seamless integration with existing logistics software.

Train and fine-tune machine learning models for document recognition and predictive analytics.
Optimize system performance for scalability and efficiency.


Requirements:

Proven experience in AI/ML development, preferably in document processing or logistics.

Proficiency in Python, TensorFlow, PyTorch, OpenCV, and NLP libraries (spaCy, NLTK, etc.).

Experience with OCR technologies (Tesseract, AWS Textract, Google Vision API).

Strong knowledge of database management (SQL, NoSQL) and cloud platforms (AWS, GCP, Azure).

Familiarity with API development (REST, GraphQL) and microservices architecture.
Excellent problem-solving skills and ability to deliver a functional MVP within a defined timeline.


This is a great opportunity to build a high-impact AI-driven solution in the logistics sector. If you’re up for the challenge, let’s talk!

Skills: Python, Artificial Intelligence, Machine Learning, TensorFlow

This project involves developing a robust and user-friendly Android app that captures or uploads images of invoices, extracts data using OCR techniques, and exports the extracted text to an Excel file in CSV format. The app should support English language only, handle both printed and handwritten text formats, and ensure high accuracy in extraction.

Key Features

1. Image Capture & Upload:

•I can capture invoice images via the camera or upload them from the gallery
•Multi-image functionality means upload or capture images not more than five images.

2. Image Preprocessing:

Apply noise removal, contrast enhancement, and other preprocessing techniques to improve OCR accuracy using tools like OpenCV.

3. OCR Text Extraction:

•Use advanced OCR SDKs such as Google ML Kit, Tesseract, or File stack etc.. for extracting invoice details like invoice number, date, item details, amounts, and vendor details etc..
•Handle both printed and handwritten text formats.

4. Error Handling:

•Notify to retake poor-quality images.
•Provide options for manual text entry if OCR fails.
•Highlight handwritten text that requires clarification.

5. Export to Excel (CSV Format):

•Organize extracted data into a predefined format provided in attached files.
•Save the Excel file locally in CSV format after only successful extraction.

Notifications & User Feedback:

•Inform users about issues like poor image quality.
•Offer retry options for capturing better images.
•If the visibility of the text in the middle of the image is poor, it is essential to address this issue before proceeding with OCR. To ensure optimal results, I recommend retaking the image with a focus on enhancing clarity and legibility.

Local Storage:

Store processed images and the generated Excel file only after successful OCR extraction on the local device.

Skills Required:

•Android app development
•OCR integration (Google ML Kit, Tesseract, File stack etc..)
•Image preprocessing (noise removal, enhancement)
•File handling (Excel generation, local storage)
•Error handling and user notifications
•Invoice data extraction and processing

Note:

•Please note that batch processing of invoices is not required, instead, we will process one invoice with multiple pages. If at all more invoices present reject it.
•invoices will be given before the project to check
•sample of CSV given in attachment
•only after showing OCR TO CSV convertion you will be awarded the project

Skills: User Interface / IA, Machine Learning (ML), OCR, Image Processing, Android App Development

An open-source AI workflow for graphic art analysis
Approach:
Combining Stable Diffusion, OpenCV, and a deep-learning model - such as TensorFlow/PyTorch -- would be a great approach.

1. Analysis Goals
Before selecting tools, determine what aspects of the artwork you want to analyze. Some possibilities include:
• Color Theory Analysis: Hue, saturation, contrast, etc.
• Composition & Layout: Rule of thirds, balance, symmetry
• Texture & Pattern Recognition; Brushstrokes, line work
• Art Style Classification: Identifying genres, e.g., Impressionism, Cubism
• Object & Subject Recognition (Detecting elements within the art)
________________________________________
2. Open-Source Tools
You will need a combination of computer vision, machine learning models, and pre-trained datasets.
A. Image Processing - OpenCV
• OpenCV will help with analyzing composition, edges, colors, and symmetry.
• Example tasks: Detecting leading lines, contrast, and contours.
B. Deep Learning - TensorFlow/PyTorch
• Use CNNs; Convolutional Neural Networks; to recognize patterns and classify artistic styles.
• Fine-tune an existing model (like ResNet or VGG) for art-specific analysis.
C. Style Recognition - Stable Diffusion & DeepDream
• Stable Diffusion can help analyze stylistic features by comparing with known datasets.
• DeepDream can highlight hidden patterns within the artwork.
________________________________________
3. Build the AI Model
Step 1: Environment
You'll need:
• Python - v3.8+
• OpenCV; pip install opencv-python
• TensorFlow or PyTorch (pip install tensorflow or pip install torch torchvision
• NumPy & Matplotlib for visualizations
Step 2: Load & Preprocess an Image
Step 3: Extract Key Features
Step 4: Train or Use a Pretrained CNN Model for Style

Skills: Artificial Intelligence, Python, Machine Learning, TensorFlow, Neural Network, Natural Language Processing, Artificial Neural Network, Graphic Design

We are developing a flagship web application for AI-powered CCTV camera analytics. This platform will serve a wide range of industries—such as retail, security, industrial, and corporate—by providing robust, real-time insights and custom analytics capabilities. The system will combine a suite of pre-built analytics with a flexible custom AI training module, enabling users to both deploy industry-standard solutions and tailor the platform to unique, evolving scenarios.

Key Objectives:
• Real-Time Analytics: Deliver live video monitoring with AI overlays (e.g., theft detection, queue monitoring, people counting, heat maps, employee productivity tracking).
• Customizability: Empower users to train and deploy their own AI models for scenario-specific analytics.
• User-Friendly Experience: Provide an intuitive, responsive UI that supports drag-and-drop dashboard customization, natural language query search, and interactive video analysis.
Functional Requirements:
• Pre-Built Analytics Modules:
• Theft/Intrusion Detection: Automated detection of suspicious behavior (e.g., shoplifting, trespassing).
• Queue Monitoring: Measure queue lengths and waiting times.
• People Counting & Heat Mapping: Real-time counting, movement tracking, and density visualization.
• Employee Productivity Tracking: Monitor employee presence/absence and zone-specific movement.
• Restricted Zone Management: Allow users to define zones and trigger alerts on breaches.
• Facial Recognition: Enable identification and verification of individuals in the footage, supporting security and access control use cases.
• Custom AI Training Module:
• Data Upload & Annotation: Users can upload video frames/images and annotate objects/events.
• Model Training & Tuning: Provide an interface for training, fine-tuning, and deploying custom AI models.
• Template Library: Offer industry-specific training templates to accelerate model creation.
• Natural Language Search:
• Query Interface: Enable users to search recorded video footage using plain language (e.g., “man in a blue hat”).
• Intelligent Filtering: Integrate NLP to process queries and return relevant video segments with context and thumbnails.
• Live Monitoring and Video Playback:
• Multi-Camera Live Feed: Support simultaneous streaming of multiple camera feeds with interactive overlays.
• Interactive Zone Drawing: Tools to draw and configure zones directly on live or recorded video.
• Event Playback: Ability to review flagged events with timeline navigation and tagging.
• Alerts and Notifications:
• Real-Time Alerts: Trigger notifications (via email, SMS, or in-app) for events such as intrusion, overcrowding, abnormal behavior, or facial recognition matches.
• Customizable Thresholds: Allow users to set sensitivity levels and parameters for alerts.
• Dashboard & Widget Customization:
• Widget Library: Pre-built widgets for live feeds, graphs, KPIs, and reports.
• Drag-and-Drop Interface: Enable users to customize dashboard layout, add/remove widgets, and reposition them.
• Reporting Tools: Interactive reports and export options (PDF, CSV) for historical data and trends.
• Admin and Client Management Portal:
• Role-Based Access Control: Different access levels for administrators, clients, and operators.
• Feature Toggling: Ability to enable/disable specific analytics modules per client.
• Audit Logs & System Monitoring: Detailed logs of system activities and model performance metrics.
Non-Functional Requirements:
• Scalability:
• Cloud-based backend (with options for on-premise deployment) to handle high video throughput and real-time processing.
• Modular microservices architecture for adding new features without affecting system performance.
• Performance:
• Low latency data processing and real-time analytics.
• Responsive UI optimized for desktop, tablet, and mobile devices.
• Security and Compliance:
• Secure user authentication and role-based access.
• Data encryption in transit and at rest.
• Compliance with relevant data privacy regulations (e.g., GDPR).
• Integration and API Support:
• RESTful APIs and WebSocket support for third-party integrations (e.g., camera systems, external AI modules).
• Integration with existing video management systems and IoT devices.
• Extensibility:
• Modular design to add future analytics features and custom integrations.
• Ability to support multiple AI models and training templates for diverse scenarios.

Skills: Computer Vision, Artificial Intelligence, TensorFlow, Deep Learning, OpenCV, Neural Network

We’re a small, innovative team building a no-code computer vision platform that lets users upload data, annotate it, train models, and (soon) deploy them—all without coding. We’re also exploring a marketplace to connect CV talent with small businesses needing AI solutions. We’re looking for a skilled computer vision freelancer to join us for a paid consulting project with a twist: beyond advising us, we want someone excited to build models on our platform and potentially use it long-term.

What You’ll Do:

Test our platform’s current features (data upload, annotation, model training) and provide feedback to improve usability and functionality.
Share insights on what CV professionals need in a no-code tool, based on your experience.
Help us refine our vision for a CV marketplace—think Fiverr or Upwork, but tailored to AI solutions.
Optionally (but ideally!), build a sample model or two using our platform to push its limits and showcase its potential.

What We’re Looking For:

Proven experience in computer vision (e.g., object detection, image segmentation, classification).
Familiarity with CV frameworks/tools (e.g., TensorFlow, PyTorch, OpenCV, etc.).
Interest in no-code solutions and simplifying AI workflows.
A proactive mindset—someone who’d enjoy testing a new tool and sticking around to create with it if it clicks.
Bonus: Experience freelancing or working with small businesses on AI projects.


Details:
Duration: 5-15 hours initially (flexible, based on your input).
Rate: $30-50/hour, depending on experience (open to negotiation).
Timeline: Starting ASAP,
Future Potential: If you love the platform, we’d welcome you to keep building models on it—maybe even feature them in our upcoming marketplace.

Skills: Python, Artificial Intelligence, Machine Learning, Web Development, Computer Vision, Internet of Things

I'm looking for an expert in Computer Vision, specifically in training models for object detection. The model will be processing real-world photos.

Ideal skills for this project include:
- Proficiency in Computer Vision techniques
- Experience in training Object Detection models
- Familiarity with real-world image processing
- Strong programming skills, preferably in Python
- Knowledge of relevant libraries such as OpenCV, TensorFlow or PyTorch

Your role will involve:
- Training the AI model to accurately detect objects in various real-world scenarios
- Ensuring the model's performance is optimised
- Providing regular updates on progress and findings

Skills: Python, Engineering, Software Architecture, CUDA, Machine Learning (ML)

We are seeking an experienced AI engineer to develop a face detection model using OpenCV. The objective is to create a robust system that validates faces against their respective IDs. The ideal candidate should have a solid understanding of computer vision and AI principles. If you have a track record of successfully implementing similar projects and are proficient in OpenCV, we want to hear from you! Please provide examples of your previous work related to face detection.

Skills: Machine Learning, Python, TensorFlow, Artificial Intelligence, Computer Vision

We are seeking an experienced consultant to provide expert guidance on implementing YOLO (You Only Look Once) and other object detection techniques for tracking individuals in video streams. The ideal candidate will assist in optimizing algorithms for real-time detection and provide insights on best practices. Your expertise will be invaluable in helping us enhance our video analysis capabilities.

If you have a strong background in computer vision, specifically with YOLO, we would love to hear from you!

Skills: YOLO, Computer Vision, OpenCV, Machine Learning, TensorFlow

Job Description: We are looking for an experienced AI/ML Developer to build a program that autom ates the processing of scanned and emailed documents and seamlessly integrates the extracted data into our CRM.

Our Firm currently processes a large volume of physical and digital documents, including client letters invoices, court orders and medical record. There documents are manually sorted, scanned and uploaded, which is time consuming. We aim to automate this process with an AI-Powered platform that can analyze, sort and process documents in the real time,

Project requirement:

Develop an AI-Powered document processing and data extraction toll.

Process both scanned physical and PDG email attachment.

Automatically extract relevant data and map iit to the correct CRM field via API

Ensure accurate OCR (Optical character recognition) and Data validation.

Enable document categorization, sorting and tagging.

Provide real time document upload to the respective client’s CRM account.

Ideal Candidate:

Expertise in AI/ML, OCR, NLP, and automation.

Proficiency in Python, TensorFlow, OpenCV or similar AI frameworks.

Experience working with document processing API’s (e.g AWS Textract, Google Cloud Vision)

Familiarity with CRM API integration (e.g. Salesforce, HubSpot, Zoho)

Cloud-based AI services (AWS, Google AI, Azure AI)



Project Scope :

AI Model development and testing.

Intégration with CRM via API

System testing and refinement.

Skills: Artificial Intelligence, Document Processing Software, NLP Tokenization, Python, Automation, API Integration, API

⸻

Title:
Build a Custom AI Model to Analyze Exterior Property Images for Signs of Vacancy or Disrepair

Description:
I am seeking a skilled AI/ML developer or team to help build a model that can analyze tens of thousands of property images (front-facing, exterior only—sourced from Google Street View) and determine whether a property appears to be:
• Vacant
• Abandoned
• In disrepair or significant need of maintenance

I will provide a large dataset of images (upwards of 50,000+) that need to be processed. The ultimate goal is to develop a system that can reliably assess these images and return a classification or confidence score for each property based on its visual condition.

Scope of Work:
• Review the dataset and assist in organizing/tagging a subset for training (if necessary)
• Build or fine-tune a computer vision model capable of identifying visual indicators of vacancy, abandonment, or disrepair
• Use machine learning (likely CNNs or other vision-based architectures) to train the model
• Evaluate and test accuracy against manually labeled images
• Deploy a simple interface or script that allows batch processing of new images
• Provide documentation and recommendations for future retraining and scaling

Ideal Skills:
• Computer vision and deep learning (TensorFlow, PyTorch, OpenCV)
• Image classification experience
• Dataset labeling/augmentation
• Model deployment (optional but helpful)
• Familiarity with Google Street View images or similar data
• Experience working with real estate or property datasets is a bonus

Deliverables:
• A trained AI model that can classify images into the appropriate categories
• Script or tool to batch process and return results (e.g., CSV with image filename + result)
• Accuracy report and documentation
• (Optional) Suggestions for improving performance or scaling for larger datasets

Additional Info:
• I am open to suggestions on model architecture or workflow improvements
• The project has strong potential for ongoing work if the solution is effective
• Budget is flexible depending on experience and proposed approach

Skills: Python, Machine Learning, TensorFlow, Artificial Intelligence, Data Science

We are seeking a skilled computer vision developer to help create a model that can accurately track objects in video footage using 3D coordinates. The ideal candidate will possess a strong background in computer vision and machine learning techniques. You will be responsible for designing, implementing, and optimizing the model to ensure precise tracking performance. Familiarity with relevant libraries and frameworks is essential. If you are a problem-solver with a passion for innovative technology, we would love to hear from you!

Skills: Computer Vision, OpenCV, Python, TensorFlow

We are seeking a skilled developer to help integrate our existing iOS app, which utilizes OpenCV for image recognition, into a React Native environment. The ideal candidate will have experience with both iOS development and React Native frameworks. You will be responsible for ensuring smooth functionality and performance of the image recognition feature within the React Native app. This project requires a strong understanding of both OpenCV and React Native, along with effective problem-solving skills. The deliverable should be a simple react native project with one screen showing the camera, and if recognized a text showing result will appear.

Skills: iOS, React Native, Smartphone, Android, Swift

**Title:** Looking for a Developer to Build an AI-Powered RTSP Camera Analysis Program

**Project Description:**
We are seeking an experienced developer to create a program that connects to an RTSP-enabled camera and analyzes player performance using AI. The system should:

- **RTSP Camera Integration** – The program should connect to an RTSP stream from a camera and process the video feed in real-time.
- **AI-Based Scoring Analysis** – Use AI to detect and track players’ shots, determining:
- Who took the shot
- How many points were scored
- Where on the target the shot landed
- **Data Logging & Visualization** – Store results and provide a dashboard or exportable report.
- **Multi-Player Support** – Detect and differentiate between multiple players.
- **Accuracy & Calibration** – Allow calibration for different camera angles and target types to ensure precise analysis.

**Requirements:**
- Experience with computer vision (OpenCV, TensorFlow, or similar)
- Familiarity with AI-based object detection and tracking
- Proficiency in working with RTSP streams and real-time processing
- Strong programming skills (Python, C++, or similar)
- Experience with UI/dashboard development is a plus

If you have experience with similar AI and video analysis projects, please send us your portfolio, estimated cost, and timeline for development.

Skills: Python

We are looking to add new features to our iOS application, specifically related to image and PDF processing.

We are working with SwiftUI, UIKit, Core Graphics, Core Image, Vision, OpenCV, and PDFKit. If you have experience with these technologies, especially in text rendering, image manipulation, PDF editing, and performance optimization, we’d love to hear from you!

📌 If you have relevant experience, please share an example of your work!

Skills: UIKit, SwiftUI, iOS, Swift, iOS Development, Objective-C

📄 Job Description:
I’m looking for a developer who can set up and host a marker-based WebAR experience, built in Unity WebGL, that runs directly in a mobile browser — no QR codes, no app download.

📌 Project Overview:
I already have a completed Unity WebGL build and all related files:
index.html, Build/ folder, JS camera tracking scripts
Image tracker logic (OpenCV, ARCamera.js, etc.)
I also have a final AR animation: a spaceship flying out of a black hole and ascending.
The system is designed for marker-based triggering using a printed design on a shirt (not QR code).

✅ Final User Flow (What You’re Building):
User visits a URL (e.g., ar.galacticquest.com) on their phone browser.
Browser requests camera access.
User points camera at a specific T-shirt design.

When detected, the animation plays:
Black hole swirls
Spaceship emerges
Ship hovers briefly
Then flies upward out of view

🎯 Deliverables:
Host all files on a secure domain or subdomain (e.g., ar.mydomain.com)
Ensure proper HTTPS configuration so mobile browsers allow camera use
Test across Safari (iOS), Chrome (Android), and ideally desktop
Provide a working link and instructions to embed/link from Shopify store
(Optional) Help optimize loading speed and camera UX

🛠 Tech Involved:
Unity WebGL
JavaScript (custom AR scripts using OpenCV or similar)
Marker/image tracking
HTTPS hosting

🔐 What I’ll Provide:
All Unity build files (.data.unityweb, .wasm, .framework.js, .loader.js)
index.html + AR JS scripts (arcamera.js, itracker.js, opencv.js)
Marker image
SSL certificate files (optional, if needed)

🧠 You Must Have Experience In:
Hosting Unity WebGL projects (not just Unity dev)
WebAR using image tracking (not QR-based)
Mobile-first browser compatibility
Deploying to a secure HTTPS server

💬 To Apply:
Please share:
1–2 examples of WebAR projects you’ve hosted/integrated before
What hosting setup you’d recommend for best performance
Your expected turnaround time

Looking to get started ASAP — thank you!

Skills: Unity, C#, Android, Game Development