Metadata-Version: 2.4
Name: teody-vision
Version: 1.0.0
Summary: A deterministic, non-neural framework for high-speed UI automation via anisotropic coherence and structure tensors.
Author-email: "Teody A. B. Banagan" <teody882011@gmail.com>
License: MIT
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Topic :: Scientific/Engineering :: Image Processing
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: numpy>=1.20.0
Requires-Dist: opencv-python>=4.5.0
Requires-Dist: mss>=6.0.0
Dynamic: license-file

# TEODY-Vision: A Deterministic, Non-Neural Framework for High-Speed UI Automation via Anisotropic Coherence and Multi-Scale Gradient Pyramids

**Author:** Teody A. B. Banagan  
**Affiliation:** University of San Carlos & Concentrix Corporation  

## Abstract
Modern Robotic Process Automation (RPA) and graphical user interface (GUI) navigation heavily rely on resource-intensive neural networks or large Vision-Language Models (VLMs). While flexible, these architectures introduce computational latency, non-deterministic inference behaviors, and heavy hardware footprints. **TEODY-Vision** is a fully deterministic, non-neural framework for high-speed UI element localization and text tracking. By combining structure tensors, anisotropic coherence ($\Psi$), multi-scale gradient magnitude pyramids, and dynamic template synthesis, TEODY-Vision achieves sub-pixel localization accuracy with minimal CPU overhead.

## Core Features
* **Deterministic Execution:** 100% execution repeatability across identical display buffers without probabilistic drift.
* **Non-Neural & Lightweight:** Zero dependency on heavy pre-trained weights or GPU accelerators; optimized for pure CPU multi-core execution.
* **Dynamic Template Generation:** Runtime synthetic template generation (`generate_text_template`), removing external static image dependencies.
* **High-Speed Screen Capture:** High-performance display buffer ingestion via `mss`.

## Installation
```bash
pip install teody-vision


### Quick Start Code Block
```
Python
import cv2
import numpy as np
from mss import MSS
from teody_vision import TEODYEngine

# Capture monitor screen buffer
with MSS() as sct:
    monitor = sct.monitors[1]
    screenshot = sct.grab(monitor)
    frame = np.array(screenshot)
    frame = cv2.cvtColor(frame, cv2.COLOR_BGRA2BGR)

# Initialize TEODY-Vision Engine
engine = TEODYEngine()

# Run the tensor coherence and topological extraction pipeline
coherence, regions = engine.process_frame(frame)

print(f"[+] Isolated {len(regions)} topological UI regions successfully.")
```

## ⚖️ Legal Manifest & Copyright Notice

* **Official Registration:** Prepared for filing under the Intellectual Property Office of the Philippines (IPOPHL) within the Computer Programs / Software category.
* **Copyright Notice:** `Copyright (c) 2026 Teody Arboneda Banagan. All rights reserved.`
* **Open Source License:** Distributed globally under the highly permissive **MIT License**.

---

## 🔬 About the Author & Project Maintainer

**Teody Arboneda Banagan, PhD (Candidate)**
*Solution Architect, AI Engineer and Researcher*
* **Email:** teody882011@gmail.com

### Professional Stature & Domain Expertise
The structural maintenance, systems design, and algorithmic orchestration of this framework are driven by 20 years of progressive IT experience and top-tier credentials spanning international enterprise teams:
* **Doctor of Philosophy (PhD) in IT Candidate** — University of San Carlos, specializing in Advanced IT Frameworks, Robotics Process Automation, Computer Vision, and Enterprise AI Architectures.
* **Master in Information Technology (MIT)** — Cebu Institute of Technology - University.
* **Peer-Reviewed Reviewer & Author** — Invited Peer Reviewer for Scopus-indexed and Web of Science international journals.
