Hello everyone,
I wanted to share an update regarding the new accessible photography
feature in PiccyBot. As we have observed, while the feature is a good
start, it currently lacks precision and creates confusion with false
directional instructions and a lack of automated features.
Yesterday, I submitted a comprehensive technical report to the PiccyBot
development team outlining a roadmap to achieve 100% accuracy and the
highest quality user experience.
The proposal covers critical improvements including "iBall" eye-tracking,
automatic capture, environmental analysis (lighting/blur checks), and full
integration for both front and back cameras to ensure we can take photos
independently without any assistance.
I am sharing the full text of the report below so you are aware of the
specific enhancements I have advocated for on behalf of our community.
Here is the full report.
PiccyBot Accessible Photography: Feature Analysis and Enhancement Roadmap
Created by: Yatin Panchal
Date: November 16, 2025
Executive Summary
The recently introduced accessible photography feature is a promising step
forward. However, current testing reveals significant discrepancies in
guidance, false positives, and a lack of critical feedback (lighting,
facial expression) that prevents users from taking photos independently.
This report details a comprehensive roadmap for perfecting the front-facing
selfie feature and extends this same logic to the back camera, enabling
users to independently photograph themselves or others with 100% accuracy
and confidence.
1. Correction of "False Positive" Directional Guidance
2. Current Issue:
When the feature is active, the app provides continuous directional prompts
(e.g., "top right center," "bottom left") even when the phone is lying flat
on a desk or facing a blank wall. This creates confusion and mistrust in
the app's reliability.
Required Fix:
Silence in Idle State: If no face is detected, the app must stop giving
directional coordinates immediately. It should instead state: "No face
detected. Please raise the device."
Threshold Implementation: Directional audio guides should only activate
when a face is identified within the frame with at least 50% confidence.
2. Advanced Facial & Gaze Analysis ("iBall" Tracking)
3. Current Issue:
The app currently centers the face but ignores facial details. A photo is
useless if the user is blinking, looking away, or not smiling.
Required Fix:
Eye Tracking (iBall System): The app must detect if eyes are open and
looking at the lens. Prompt: "Look slightly up at the camera" or "Open your
eyes wider."
Expression Detection: The app must detect the mouth shape. Prompt: "You are
not smiling," or "Perfect smile detected."
Centering Accuracy: Instructions must be granular (e.g., "Move phone
slightly left" rather than just "Left").
3. Environmental & Quality Assessment
4. Current Issue:
Users are currently taking photos that are centered but unusable due to
blurriness, darkness, bad backgrounds, incorrect distance, or an obstructed
lens.
Required Fix:
Proximity Guidance: The app must detect the face's distance. Prompt: "Move
phone further away" or "Move phone closer for a better shot."
Lens Obstruction: The app must detect if a finger or object is blocking the
camera. Prompt: "Camera lens is blocked."
Lighting Analysis: Real-time feedback on illumination. Prompt: "Too dark,
please find a light source," or "Strong light behind you, please turn
around."
Blur Detection: If the autofocus is hunting or the user’s hand is shaking,
the app must warn the user. Prompt: "Image is blurry, hold steady."
Background Aesthetics: AI should analyze the background scene. Prompt:
"Background is cluttered," or "Background is clear."
4. Implementation of "Auto-Capture" (Crucial Update)
5. Current Issue:
Prompting a blind user to manually press a shutter button causes the phone
to shake, shifting the framing at the last second and ruining the shot.
Required Fix:
Automatic Shutter Trigger: The app must take the photo automatically only
when all conditions are met:
1. Face(s) are centered.
2. Eyes are open and looking at the lens.
3. Subject(s) are smiling (optional setting).
4. Image is in focus (not blurry).
5. Lighting is adequate.
6. Proximity is correct.
7. Countdown: Once conditions are met, a short audio countdown (e.g., "3...
2... 1... [Shutter Sound]") allows the user to freeze their pose.
8. 5. Back Camera Integration (Group Photography & High-Res Selfies)
9. Current Issue:
The current accessibility logic is limited to the front camera. Users
cannot take high-quality selfies with the back camera, nor can they
reliably take photos of other people or groups.
Required Fix:
Mode Extension: All detection logic (Face, Gaze, Expression, Lighting,
Blur, Proximity) must be extended to the back camera.
Face Counting: The app must announce the number of faces detected. Prompt:
"One face detected," or "Three faces detected."
Group Centering: Guidance must adapt to frame all subjects. Prompt: "Move
phone left to center the group," or "One person is cut off on the right."
Group Status: The app should check the status of all faces. Prompt:
"Someone's eyes are closed," or "All faces clear. Hold steady."
This integration enables two key use cases: taking superior quality selfies
with the main camera and independently photographing other people.
6. Post-Capture Workflow & Navigation
7. Current Issue:
There is no logical flow after taking the photo. Users are unsure if the
photo is saved or how to get a description.
Required Fix:
Immediate Post-Capture Menu: Immediately after the auto-capture, the app
must pause and ask: "Photo taken. Double tap to Save to Gallery, or swipe
to Describe."
The "Describe then Save" Loop: If the user chooses "Describe," the app must
generate the description. Crucially, after the description is read, the app
must ask again: "Would you like to save this photo to the gallery?"
Reasoning: A user cannot decide to save a photo until they know what it
looks like.
7. Additional Technical Recommendations for 100% Accuracy
8. Haptic Feedback (Vibration):
Use vibration patterns alongside voice commands. A fast vibration indicates
"You are close," and a solid, continuous vibration indicates "Perfect
alignment." This helps in noisy environments.
Phone Tilt/Angle Guidance:
Often, a user centers their face, but the phone is tilted forward or
backward (pitch/roll). Prompt: "Tilt the top of the phone away from you" or
"Hold the phone upright."
Burst Mode Selection:
To ensure the "highest quality," the app should silently take a burst of 3
photos and use AI to automatically save the one with the sharpest focus and
best smile.
Conclusion
By moving from simple "face centering" to a holistic "Smart Photography
Assistant," PiccyBot can become the definitive and all-encompassing tool
for accessible photography. Implementing this complete roadmap—covering
both front and back cameras, environmental checks, and a logical
auto-capture workflow—will empower users to confidently capture both
selfies and photos of the world around them.
Thanks and Regards,
Yatin Panchal
B.A. in English Literature, University of Gujarat, Class of 2019.
Writer, Poet, Lyricist | Advocate for Inclusivity and Accessibility.
Certification: Certified NVDA Screen Reader Expert, January 2022
AI and Cybersecurity Enthusiast
"Inspiring inclusivity and empowerment through innovation and words."
Contact Information:
E-mail: [email protected]
Facebook: Yatin Panchal
Instagram: yatin.panchal143
Twitter: yatinpanchal143
Snapchat: yatinpanchal143
Telegram: @glamorous143
Gujarat, India.
--
Disclaimer:
1. Contents of the mails, factual, or otherwise, reflect the thinking of the
person sending the mail and AI in no way relates itself to its veracity;
2. AI cannot be held liable for any commission/omission based on the mails sent
through this mailing list..
Search for old postings at:
http://www.mail-archive.com/[email protected]/
---
You received this message because you are subscribed to the Google Groups
"AccessIndia" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion visit
https://groups.google.com/a/accessindia.org.in/d/msgid/accessindia/CALN9kYJT39mmp-Dw6TO27Rf3opc%2BNj8LVwLvEQsXetEynnrRZQ%40mail.gmail.com.
