ComfyUI Node: ImageQualityScoreNode

ImageQualityScoreNode:

The ImageQualityScoreNode is designed to evaluate and score the quality of an image based on various criteria, providing a comprehensive assessment that can be used to enhance image processing workflows. This node calculates a total quality score by considering multiple factors such as aesthetic appeal, the presence of desirable and undesirable features, and the likelihood of the image being artificially generated or human-made. By assigning weights to each of these factors, the node allows for a customizable scoring system that can be tailored to specific needs or preferences. The primary goal of this node is to offer a nuanced and flexible approach to image quality evaluation, enabling users to make informed decisions about image selection and processing.

ImageQualityScoreNode Input Parameters:

aesthetic_score

The aesthetic_score parameter represents the perceived aesthetic value of the image. It is a floating-point number that contributes to the overall quality score, with higher values indicating more aesthetically pleasing images. This parameter is multiplied by the weight_aesthetic_score to determine its impact on the final score.

image_score_good

The image_score_good parameter quantifies the presence of desirable features in the image. It is a floating-point number that, when multiplied by the weight_good_score, adds to the total quality score. A higher value suggests that the image contains more positive attributes.

image_score_bad

The image_score_bad parameter measures the presence of undesirable features in the image. It is a floating-point number that, when multiplied by the weight_bad_score, detracts from the total quality score. A higher value indicates more negative attributes in the image.

ai_score_artificial

The ai_score_artificial parameter assesses the likelihood of the image being artificially generated. It is a floating-point number that, when multiplied by the weight_AIDetection, reduces the total quality score. A higher value suggests a greater probability of the image being AI-generated.

ai_score_human

The ai_score_human parameter evaluates the likelihood of the image being human-made. It is a floating-point number that, when multiplied by the weight_HumanDetection, adds to the total quality score. A higher value indicates a greater probability of the image being created by a human.

weight_good_score

The weight_good_score parameter is a multiplier that determines the influence of the image_score_good on the total quality score. It allows users to adjust the importance of desirable features in the scoring process.

weight_aesthetic_score

The weight_aesthetic_score parameter is a multiplier that determines the influence of the aesthetic_score on the total quality score. It allows users to adjust the importance of aesthetic appeal in the scoring process.

weight_bad_score

The weight_bad_score parameter is a multiplier that determines the influence of the image_score_bad on the total quality score. It allows users to adjust the importance of undesirable features in the scoring process.

weight_AIDetection

The weight_AIDetection parameter is a multiplier that determines the influence of the ai_score_artificial on the total quality score. It allows users to adjust the importance of AI detection in the scoring process.

weight_HumanDetection

The weight_HumanDetection parameter is a multiplier that determines the influence of the ai_score_human on the total quality score. It allows users to adjust the importance of human detection in the scoring process.

MultiplyScoreBy

The MultiplyScoreBy parameter is a multiplier applied to the final calculated score. It allows users to scale the total quality score to a desired range or magnitude.

show_on_node

The show_on_node parameter is a boolean that determines whether the calculated score and its components are displayed on the node. If set to True, the score details are shown, providing transparency and insight into the scoring process.

ImageQualityScoreNode Output Parameters:

result

The result parameter is a tuple containing the final calculated quality score as a floating-point number. This score represents the overall quality of the image based on the weighted contributions of the input parameters.

ui

The ui parameter is a dictionary that contains the formatted score details if show_on_node is set to True. It provides a textual representation of the score and its components, offering users a clear understanding of how the final score was derived.

ImageQualityScoreNode Usage Tips:

• Adjust the weights of the scoring parameters to emphasize the aspects of image quality that are most important for your specific application, such as aesthetics or human-like features.
• Use the MultiplyScoreBy parameter to scale the final score to a range that aligns with your evaluation criteria or integrates seamlessly with other scoring systems.

ImageQualityScoreNode Common Errors and Solutions:

Error calculating score: <error_message>

• Explanation: This error occurs when there is an issue with the input values or the calculation process, such as invalid data types or division by zero.
• Solution: Ensure that all input parameters are provided with valid numerical values and that the weights are set appropriately. Check for any potential division by zero in the calculation process.