Thai Entrance Face Image Dataset
Budget: $250 – $750 USD
Hi there — I’m ABD from Saytica.
A quick bit of context before the specs: our team builds real-world training datasets for computer vision and AI teams. We don’t “just collect files”—we design a clean workflow (consent → capture → QA → delivery) so the final dataset is usable immediately, without rework.
Right now, we’re launching a Thailand offline face-image collection and we’re looking for a Thailand-based partner who can run the full operation end-to-end: recruitment + venue + field crew + equipment + on-site quality control. The goal is to finish within one month.
Why this matters: small mistakes in setup (camera angle, blur, inconsistent lighting, missing metadata) can ruin a dataset. So we’re being very specific upfront, and we’ll move fast with the right team. The goal is to photograph 200 Thai nationals as they walk through shop or hall-lobby entrances; renting an appropriate venue is perfectly fine as long as ceiling-mounted cameras look straight toward the door.
Camera & capture setup
• One or more ceiling-mounted cameras positioned so the lens faces the entrance
• Frontal view only, with yaw up to ±45° and pitch up to ±30° captured at far, medium and near distances
• Native resolution must be 2304 × 1296 px; no up- or down-scaling
• Please keep lighting consistent and avoid motion blur—participants should appear clearly in every frame
Sampling requirements
• 1,670 Thai participants
• Age mix: 10 % under 18, 30 % ages 18-30, 30 % ages 30-40, 30 % over 40
• An even gender split overall
• 15 images per person (one per angle per distance), totalling 25,050 photos
Deliverables & acceptance criteria
1. Folder structure that groups images by participant ID.
2. Raw, uncompressed files at the specified resolution.
3. CSV manifest listing participant ID, age bracket, gender, angle label and distance label for every file.
4. A small, watermarked sample set for my approval after the first 50 participants.
5. Final dataset delivered on an encrypted drive or secure cloud link once everything passes a visual quality check.
If you already have ceiling-mountable hardware and a workflow for large-scale face data capture, you’ll hit the ground running. Let’s discuss scheduling, team size, and any on-site logistics you need from me so we can release the first sample set right away.
A quick bit of context before the specs: our team builds real-world training datasets for computer vision and AI teams. We don’t “just collect files”—we design a clean workflow (consent → capture → QA → delivery) so the final dataset is usable immediately, without rework.
Right now, we’re launching a Thailand offline face-image collection and we’re looking for a Thailand-based partner who can run the full operation end-to-end: recruitment + venue + field crew + equipment + on-site quality control. The goal is to finish within one month.
Why this matters: small mistakes in setup (camera angle, blur, inconsistent lighting, missing metadata) can ruin a dataset. So we’re being very specific upfront, and we’ll move fast with the right team. The goal is to photograph 200 Thai nationals as they walk through shop or hall-lobby entrances; renting an appropriate venue is perfectly fine as long as ceiling-mounted cameras look straight toward the door.
Camera & capture setup
• One or more ceiling-mounted cameras positioned so the lens faces the entrance
• Frontal view only, with yaw up to ±45° and pitch up to ±30° captured at far, medium and near distances
• Native resolution must be 2304 × 1296 px; no up- or down-scaling
• Please keep lighting consistent and avoid motion blur—participants should appear clearly in every frame
Sampling requirements
• 1,670 Thai participants
• Age mix: 10 % under 18, 30 % ages 18-30, 30 % ages 30-40, 30 % over 40
• An even gender split overall
• 15 images per person (one per angle per distance), totalling 25,050 photos
Deliverables & acceptance criteria
1. Folder structure that groups images by participant ID.
2. Raw, uncompressed files at the specified resolution.
3. CSV manifest listing participant ID, age bracket, gender, angle label and distance label for every file.
4. A small, watermarked sample set for my approval after the first 50 participants.
5. Final dataset delivered on an encrypted drive or secure cloud link once everything passes a visual quality check.
If you already have ceiling-mountable hardware and a workflow for large-scale face data capture, you’ll hit the ground running. Let’s discuss scheduling, team size, and any on-site logistics you need from me so we can release the first sample set right away.