Detectify

Inspiration

Many of our team members have family or friends who struggle with vision impairment. This inspired us to create a solution that could assist them and others facing similar challenges, helping ensure safety and independence even if vision declines in the future.

What It Does

Our AI leverages the device’s local camera to detect people, vehicles, stop signs, traffic lights, and crosswalks. It identifies whether objects are moving or stationary and estimates their distance from the camera. Safety checks alert the user when an object is too close our text-to-speech (TTS) system announces “danger,” specifies the object type, and reports its proximity.

How We Built It

We developed Detectify using the YOLOv8 API and Python in Visual Studio Code and IDLE. YOLOv8 connects to local cameras, scans images, and outlines detected objects such as people and vehicles with color-coded bounding boxes. We implemented a distance calculation formula using reference points to estimate object proximity in real time. By comparing object positions frame by frame, our system determines whether an object is approaching, moving away, or stationary.

Challenges We Faced

We encountered challenges with integrating text-to-speech smoothly, finalizing the distance formula, designing the workflow, and debugging various implementation issues.

Accomplishments We’re Proud Of

We’re proud to have built a fully functional detection system within a short timeframe. Our TTS integration performed beyond expectations, and overcoming numerous coding errors through persistence and collaboration strengthened our skills in AI, teamwork, and problem-solving.

What We Learned

We gained experience implementing APIs, automating workflows, collaborating effectively with GitHub, and understanding the real-world development process behind AI/ML projects.

What’s Next for Detectify

Our goal is to integrate Detectify into smart glasses equipped with speakers. We plan to enhance detection accuracy and speed while exploring opportunities to turn this project into a full startup. Ultimately, we hope Detectify will be adopted in the medical field and positively impact lives worldwide.

Built With

• cmd
• copilot
• gemini
• github
• idle
• javascript
• powershell
• python
• vscode
• yolo