# Changelog Livennes

**2.5**

- General optimization of the capture and validation flow to improve stability across consecutive sessions.
- Performance adjustments to reduce latency when returning results to the host.
- Initial improvements in evidence reduction (Base64) to decrease payload size.

**2.6.0**

- Camera pipeline optimization for smoother performance on mid-range devices.
- Evidence compression adjustments to reduce transfer size and improve delivery times.
- UX improvements with clearer feedback during alignment and capture.

**2.7.0**

- Facial tracking optimization for better face stability within the frame.
- Performance improvements in frame processing to reduce sustained CPU/GPU load.
- Flow logic adjustments to improve consistency between capture, analysis, and results.

**2.8.0**

- Optimization of evidence return to the host through more efficient Base64 generation and encoding.
- Further reduction of image/evidence size while maintaining reference quality.
- Visual experience improvements for greater consistency across resolutions and devices.

**2.9.0**

- Optimization of camera startup and shutdown, as well as flow recovery during retries.
- Memory usage improvements to prevent performance degradation after multiple runs.
- User feedback adjustments to reduce friction and increase process clarity.

**3.0.0**

- Major SDK evolution focused on performance and lifecycle stability of the component.
- Analysis pipeline optimization for smoother execution and more consistent response times.
- Improvements in resource handling (camera and memory) for longer sessions.

**3.0.3**

- Optimization of camera session startup and transition into analysis to reduce startup times.
- Efficiency adjustments to reduce CPU/GPU spikes during initialization.

**3.0.8**

- Improvements in evidence compression and serialization: lighter Base64 and faster delivery to the host.
- Flow optimization to reduce latency between capture and final result.

**3.1.0**

- Logic adjustments for greater consistency in the validation flow.
- Memory usage optimization in repeated sessions to improve stability.

**3.1.5**

- UX improvements with clearer and more consistent feedback during the capture process.
- Performance optimization on resource-constrained devices to maintain smoothness.

**3.2.0**

- Optimization of the retry system without full reinitialization to improve flow continuity.
- Evidence compression adjustments to reduce payload size while maintaining adequate quality.

**3.2.6**

- Optimization of frame processing to reduce sustained CPU/GPU load.
- Camera pipeline stability improvements for uninterrupted analysis.

**3.3.0**

- Increased robustness of facial analysis under lighting variations and moderate movement.
- Optimization of evidence generation and encoding for faster return to the host.
- Performance adjustments to maintain more stable FPS during the session.

**3.3.7**

- Optimization of camera module initialization time and analysis preparation.
- Internal adjustments for a more stable experience during prolonged runs.

**3.4.0**

- Facial tracking optimization for greater consistency of the face throughout the session.
- Performance improvements in evidence compression (Base64) to reduce transfer size and timing.
- Visual experience adjustments for consistency across different resolutions and devices.

**3.4.6**

- Optimization of memory usage and resource release at the end of the flow to maintain stability.
- UI/Render performance improvements through reduced style load and smoother transitions.

**3.5.0.1**

- Integration of the new Liveness engine with advanced face detection and emotion/gesture analysis.
- Optimization of the capture and validation flow.
- Design and user experience improvements (UI/UX).
- Reduced latency in the camera pipeline.
- Greater robustness against flow manipulation.

**3.5.2.0**

- Improved facial detection system for gesture recognition.
- Specific optimization for devices with Snapdragon 8 Elite.
- Strengthened anti-spoofing signals.
- Greater efficiency in resource consumption during active sessions.

**3.5.2.1**

- Optimization of component initialization and instantiation.
- Localization and language consistency improvements.
- Additional performance adjustments.
- Refinement of control logic and result delivery.