I've been actively exploring custom AI models for various tasks in OptiCull, and I'm excited to share the updates for version 1.6.0.
Reliably detect faces at longer distances
In version 1.6.0, I've incorporated a custom model that ensures reliable face detection even at longer distances. This enhancement is particularly beneficial for wedding and sports photographers, offering more accurate face detections in scenarios like group shots.
The challenge was to integrate a new model that is not only more accurate but also as fast as the current one. Additionally, the new model needed to leverage the neural engines built into Apple Silicon Macs for maintaining efficiency. I'm pleased to report that the new model is not only faster but also more efficient and accurate.
The new model, currently in beta, is not enabled by default. Take it for a spin by enabling custom AI models from the AI pane in OptiCull settings. Based on your feedback, I'll make it the default model for face detections in a future update.
With the addition of this custom model, it becomes easier to introduce models for other tasks in the future, such as blink, emotions, and focus detections. Stay tuned for more updates!
Moving to subscription only pricing
We're thrilled to announce that all early adopter slots for the Lifetime plan have been filled, and we're transitioning exclusively to subscription plans. Our subscription option remains highly affordable at just USD 4.99 per month when billed annually.
A sincere thank you for all the valuable feedback and feature requests. Your input has been instrumental in shaping my decisions for upcoming features.
Changelog
1.6.0 - Dec 5, 2023
- Introducing Custom AI models (Beta), starting with improved face detection accuracy.
- Preference to enable Custom AI models (Beta) and reanalyze old projects upon opening.