AI Vision System for Fall Detection

Development and Testing of an AI-Based Computer Vision Trip and Fall Detection System

1. Invitation to Tender

Transpix AI invites technology providers to submit proposals for the development and testing of an Artificial Intelligence (AI) and Computer Vision solution capable of detecting trips and falls in a controlled testing environment.

The project will commence on 1 Sep 2026 and conclude on 19 December 2026.

2. Project Background

The objective of this project is to develop an intelligent computer vision system that can automatically detect trip and fall events using video data. The system should utilise modern AI techniques to distinguish between normal activities and fall-related incidents while minimising false alarms.

The project will focus on algorithm development, model training, validation, and controlled environment testing.

3. Project Objectives

The successful supplier shall:

· Develop an AI-based computer vision model capable of detecting trip and fall events.

· Design and implement the complete detection pipeline.

· Train and optimise the model using appropriate datasets.

· Validate system performance.

· Test the solution in a controlled environment.

· Produce performance reports and recommendations for future deployment.

4. Scope of Work

The supplier shall undertake the following activities.

Phase 1 – Requirements and Design

· Review project requirements.

· Design the overall AI system architecture.

· Select suitable computer vision algorithms.

· Identify hardware and software requirements.

Phase 2 – AI Model Development

· Develop the trip and fall detection model.

· Implement person detection and tracking.

· Train deep learning models using suitable datasets.

· Optimise inference speed and detection accuracy.

Phase 3 – System Integration

· Integrate the AI model with video input.

· Develop alert generation capability.

· Implement event logging.

· Develop reporting functionality.

Phase 4 – Controlled Testing

Testing will be conducted within a controlled environment.

Testing shall include:

· Normal walking

· Sitting

· Standing

· Running

· Intentional trips

· Simulated falls

· Multiple people in view

· Different lighting conditions

· Camera angle variation

· Partial occlusion scenarios

Performance metrics shall include:

· Detection Accuracy

· Precision

· Recall

· Detection Latency

· System Reliability

5. Deliverables

The successful supplier shall provide:

· Project Management Plan

· System Design Documentation

· AI Model

· Source Code

· Training Documentation

· Test Plan

· Test Reports

· Validation Report

· User Guide

· Final Project Report

6. Project Schedule

Milestone Date

Project Kick-off 1 Sep 2026

Initial AI Model 30 September 2026

Model Optimisation 31 October 2026

Controlled Testing 1–30 November 2026

Final Validation 10 December 2026

Final Deliverables 19 December 2026

7. Technical Requirements

The proposed solution should:

· Use modern computer vision techniques.

· Employ machine learning and/or deep learning methods.

· Process live or recorded video streams.

· Operate in near real time.

· Detect trip and fall events with high accuracy.

· Be scalable for future deployment.

· Support common video formats.

· Produce configurable alerts and event logs.

8. Supplier Requirements

Tenderers should demonstrate:

· Experience in AI and machine learning.

· Experience in computer vision.

· Previous delivery of similar AI solutions.

9. Tender Submission Requirements

Tender submissions should include:

· Company profile

· Relevant experience

· Proposed technical approach

· Project schedule

· Project team and key personnel

· Cost proposal

10. Project Duration

Project Start: I Sep 2026

Project Completion: 19 December 2026

Total Duration: 4 months

11. Intellectual Property

All source code, trained AI models, documentation, datasets developed specifically for this project, and related intellectual property produced under this contract shall become the property of the client upon final payment unless otherwise agreed in writing.

12. Confidentiality

The successful supplier shall treat all project information, datasets, videos, software, and documentation as confidential and shall comply with all applicable data protection and privacy requirements.

13. Acceptance Criteria

The project will be deemed complete upon:

· Successful completion of all planned deliverables.

· Demonstration of the AI trip and fall detection system.

· Completion of controlled testing.

· Submission of all documentation.

· Acceptance of the final report by the client.