Computer Vision
Flunkey
Facial Activity Recognition System
Developed video processing model with hybrid VGG16/Conv2D architecture for facial expression analysis. Used TensorFlow and Albumentations for image augmentation with LabelMe annotation achieving 98% frame-level accuracy. REST API integration reduced latency by 25%, increasing customer satisfaction by 15-20% across partnered F&B outlets.
5 months
Project Duration
98% frame-level accuracy
Key Achievement
6+
Technologies Used
The Challenge
Flunkey required a facial activity recognition system for customer experience analysis that could process video streams in real-time with high accuracy.
The Solution
Developed a hybrid architecture combining VGG16 for feature extraction with custom Conv2D layers, integrated with Albumentations for data augmentation and LabelMe for annotation management.
Results & Impact
- Achieved 98% frame-level accuracy
- Reduced inference latency by 25%
- Increased customer satisfaction by 15-20%
- Processed real-time video streams efficiently
Project Metrics
98% frame-level accuracy
25% latency reduction
15-20% customer satisfaction boost
Technologies Used
TensorFlowAlbumentationsVGG16Conv2DLabelMeREST API
Project Details
Company
Flunkey
Duration
5 months
Team Size
3 engineers + 1 computer vision specialist
Category
Computer Vision