chloel0.bsky.social
@chloel0.bsky.social
VERiiO Dev Log - Week 4
After training the new CV model in Roboflow, I realised that the annotating medthod needed to be changed to cut out more unrelated areas for a more accurate result.
After training the new CV model in Roboflow, I realised that the annotating medthod needed to be changed to cut out more unrelated areas for a more accurate result.
February 5, 2026 at 5:14 PM
VERiiO Dev Log - Week 4
After training the new CV model in Roboflow, I realised that the annotating medthod needed to be changed to cut out more unrelated areas for a more accurate result.
After training the new CV model in Roboflow, I realised that the annotating medthod needed to be changed to cut out more unrelated areas for a more accurate result.
VERiiO Dev Log - Week 4
Back in Unreal this week, leveling up our synthetic data generation.
The CV model needed better training data, so I rebuilt the synth pipeline for higher quality outputs. Cleaner texture, better lighting, more realistic defects.
Back in Unreal this week, leveling up our synthetic data generation.
The CV model needed better training data, so I rebuilt the synth pipeline for higher quality outputs. Cleaner texture, better lighting, more realistic defects.
February 2, 2026 at 2:53 AM
VERiiO Dev Log - Week 4
Back in Unreal this week, leveling up our synthetic data generation.
The CV model needed better training data, so I rebuilt the synth pipeline for higher quality outputs. Cleaner texture, better lighting, more realistic defects.
Back in Unreal this week, leveling up our synthetic data generation.
The CV model needed better training data, so I rebuilt the synth pipeline for higher quality outputs. Cleaner texture, better lighting, more realistic defects.
VERiiO Dev Log - Week 3
Took a break from 3D this week to tackle the UI.
Brought in a UI designer to overhaul the user flow and nail down our branding direction.
Took a break from 3D this week to tackle the UI.
Brought in a UI designer to overhaul the user flow and nail down our branding direction.
February 2, 2026 at 2:47 AM
VERiiO Dev Log - Week 3
Took a break from 3D this week to tackle the UI.
Brought in a UI designer to overhaul the user flow and nail down our branding direction.
Took a break from 3D this week to tackle the UI.
Brought in a UI designer to overhaul the user flow and nail down our branding direction.
VERiiO Dev Log - Week 2
Last week's 3D visualization concept actually worked! But the model was bloated and didn't match industry inspection standards.
Spent this week rebuilding it from scratch—optimized and compliant
Last week's 3D visualization concept actually worked! But the model was bloated and didn't match industry inspection standards.
Spent this week rebuilding it from scratch—optimized and compliant
February 2, 2026 at 2:44 AM
VERiiO Dev Log - Week 2
Last week's 3D visualization concept actually worked! But the model was bloated and didn't match industry inspection standards.
Spent this week rebuilding it from scratch—optimized and compliant
Last week's 3D visualization concept actually worked! But the model was bloated and didn't match industry inspection standards.
Spent this week rebuilding it from scratch—optimized and compliant
VERiiO Dev Log:Week 1.
This week I dove into visualizing 3D defect reports. The goal: shift color gradients based on defect density on a 3D model.
The research was hard but after talking to a few people, I landed on an approach that might actually work.
This week I dove into visualizing 3D defect reports. The goal: shift color gradients based on defect density on a 3D model.
The research was hard but after talking to a few people, I landed on an approach that might actually work.
February 2, 2026 at 2:39 AM
VERiiO Dev Log:Week 1.
This week I dove into visualizing 3D defect reports. The goal: shift color gradients based on defect density on a 3D model.
The research was hard but after talking to a few people, I landed on an approach that might actually work.
This week I dove into visualizing 3D defect reports. The goal: shift color gradients based on defect density on a 3D model.
The research was hard but after talking to a few people, I landed on an approach that might actually work.