Dataset-ISWA_200479
Description
The data in the CSV file represents an initial collection used in the system developed for synthesizing emotional expressions using Generative Adversarial Networks (GAN). This dataset captures key metrics and features relevant to evaluating the quality of the generated images and the system's performance under different conditions. The main variables included are described below: 1. Emotion: This column categorizes the emotions generated by the system, covering four central states: Joy, Sadness, Surprise, and Anger. These emotions are critical to assess the model's ability to capture various emotional expressions. 2. SSIM (Structural Similarity Index): This index indicates the structural quality of the generated images, comparing them to authentic expressions. Values range from 0 to 1, with higher values reflecting greater visual and structural similarity. 3. LPIPS (Learned Perceptual Image Patch Similarity) evaluates the visual perception of the generated images. This indicator complements the SSIM and provides a metric more aligned with human perception. Lower values are available for high-quality photos. 4. Precision: Represents the percentage of emotions correctly generated by the system, highlighting its precision in emotional synthesis. It is measured as a percentage value that reflects the model's fidelity in emotional transmission. 5. False Positives: This indicator counts the cases in which the system incorrectly classified a non-existent emotion. It is key to analyzing the model's robustness and ability to avoid classification errors. 6. False Negativesrecords cases where the system failed to detect a present emotion. This indicator is critical to understanding the model's limitations in more complex situations. 7. CPU Usage (%): Represents the percentage of CPU usage during emotion processing, reflecting the system's efficiency in terms of computational resources. 8. GPU Usage (%): Similar to CPU usage, this variable measures the percentage of GPU usage, which is key in executing operations related to deep networks. 9. Memory Usage (MB): This indicates memory consumption during processing, evaluating the system's scalability under high-demand conditions. 10. Frame Rate (FPS): Measures the number of frames processed per second, an essential factor in maintaining a fluid real-time experience. 11. Response Time (ms): This represents the total time from data capture to the generation of the emotional expression. It is essential for evaluating the system's capacity in interactive and high-load scenarios. This data provides a comprehensive overview of system performance regarding accuracy, generation quality, resource usage, and real-time responsiveness. By analyzing them, areas of improvement and optimization of the model can be identified for applications in virtual educational environments and other interactive platforms.